The Nutritional-Toxicological Conflict related to Seafood ...fatty acids. Next, an introduction to...

The Nutritional-Toxicological Conflict related to Seafood Consumption

Faculty of Medicine and Health Sciences Department of Public Health

The Nutritional-Toxicological Conflict related to Seafood Consumption

ir. Isabelle Sioen

Thesis submitted in fulfilment of the requirements for the degree of Doctor in Medical Sciences

Promoters:

Prof. Dr. S. De Henauw (Department of Public Health – Faculty of Medicine and Health Sciences)

Prof. Dr. ir. J. Van Camp

(Department of Food Safety and Food Quality – Faculty of Bioscience Engineering)

Sioen, Isabelle The Nutritional-Toxicological Conflict related to Seafood Consumption PhD-thesis Ghent University – with references – with summary in Dutch Copyright © 2007, Isabelle Sioen All rights reserved. No part of this thesis may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording or any information storage or retrieval system, without permission in writing from the author, or, when appropriate, from the publisher of the publications. Financial support for this PhD-work: Institute for the Promotion of Innovation through Science and Technology in Flanders (IWT-Vlaanderen) The Belgian Science Policy through the SPSD II project CP/02/56 The food consumption data used in this PhD-work were financially supported by: The National Fund for Scientific Research/Flemish Division (Fund No G.0152.01 and Fund No 31557898), the Kellogg’s Benelux Company, Unilever Belgium, the Belgian Nutrition Information Center, and the Integrated Research Project SEAFOODplus, contract FOOD-CT-2004-506359 (partial financing by the EU) Verschenen in de reeks monografieën van de Vakgroep Maatschappelijke Gezondheidkunde, Universiteit Gent ISBN 9789078344070 D/2007/4531/4 Printed by DCL Print & Sign | Leegstraat 15 | 9060 Zelzate | www.dclsigns.be

“You shall offer something good to your body, so that the soul feels well to live in it.”

Anonymous

Members of the Jury Prof. Dr. S. De Henauw (promoter) Department of Public Health, Ghent University Prof. Dr. ir. J. Van Camp (promoter) Department of Food Safety and Food Quality, Ghent University Prof. Dr. J.L. Willems Department of Public Health, Ghent University Prof. Dr. G. De Backer Department of Public Health, Ghent University Prof. Dr. J.M. Kaufman Department of Internal Medicine, Ghent University Prof. Dr. J. Van de Voorde Department of General Physiology and Human Physiology and Pathophysiology, Ghent University Prof. Dr. C. Van Peteghem Department of Bio-analysis, Ghent University Prof. Dr. ir. G. Vansant Department of Public Health, Catholic University of Leuven Prof. Dr. ir. W. Verbeke Deparment of Agricultural Economics, Ghent University Dr. J.L. Volatier Direction Risk Assessment Nutrition & Food Safety, French Food Safety Agency

Abbreviations %E percentage of total energy intake AA arachidonic acid AHA American Heart Association AMDR acceptable macronutrient distribution range bw body weight CEC Central Economic Council CHD coronary heart disease COT Committee on Toxicity (United Kingdom) Cons. consumers-only CVD cardiovascular disease DHA docosahexaenoic acid dl PCBs dioxin-like PCBs DPA docosapentaenoic acid DRI dietary reference intake EDR estimated diet record EFSA European Food Safety Authority EPA eicosapentaenoic acid FA fatty acid FAO Food and Agriculture Organization FCDB food composition database FCS food consumption survey FFQ food frequency questionnaire Hg mercury HIS health interview survey ICES International Council for the Exploration of the Sea IOM Institute of Medicine (United States of America) iPCBs indicator PCBs ISSFAL International Society for the Study of Fatty Acids ans Lipids JECFA Joint FAO/WHO Expert Committee on Food Additives LA linoleic acid LNA α -linolenic acid LC long chain LC n-3 PUFAs long chain omega-3 poly-unsaturated fatty acids LOD limit of detection LOQ limit of quantification MeHg methyl mercury n-3 omega-3 n-6 omega-6 PBDEs polybrominated diphenyl ethers ndl PCBs non-dioxin like PCBs PCBs polychlorinated biphenyls PCDDs polychlorinated dibenzo-p-dioxins PCDFs polychlorinated dibenzofurans PUFAs poly-unsaturated fatty acids SACN Scientific Advisory Committee on Nutrition (United Kingdom) SCO single cell oil SFA saturated fatty acid

TDI tolerable daily intake TEFs toxic equivalency factors TEQ toxic equivalent totTEQ total TEQ TWI tolerable weekly intake UL tolerable upper intake level WHO World Health Organization

Table of contents Chapter I. General introduction Chapter II. Overall dietary PUFA intake and importance of seafood as

PUFA and vitamin D source Chapter III. Traceability and nutrient and contaminant content of seafood

on the Belgian market: elaboration of databases Chapter IV. Probabilistic intake assessment of multiple compounds as a

tool to quantify the nutritional-toxicological conflict related to seafood consumption

Chapter V. General discussion References Summary Samenvatting Dankwoord/Acknowledgments About the author Publications of the author

15

41

75

131

161

181

215

219

223

227

229

Chapter I. General introduction

15

Chapter I. General introduction Pioneer research in the nineteen sixties and seventies by Danish scientists indicated that the

consumption of fish was associated with a reduced risk for heart disease in Greenland

Eskimos (Bang & Dyerberg, 1980; Dyerberg et al, 1975). This Eskimo population experienced

a low mortality from coronary heart disease despite a diet rich in fat and cholesterol. The

investigators proposed that this could be related to the high content of omega-3 fatty acids,

typically present in marine foods.

These Eskimo studies have triggered a much broader and intensified research on the

importance of omega-3 fatty acids and seafood in the human diet. This PhD-thesis is

embedded in this research area and examines in the first place the intake of omega-3 and

other fatty acids by the Flemish and Belgian population. In a next step, the question is raised

whether seafood is a safe dietary source of these fatty acids and whether the consumption

conform with physiological needs induces any toxicological concern. The latter is of

importance since the favourable health perception is troubled by information regarding the

potential adverse health impact of chemical contaminants in marine foods, occuring

naturally or resulting from man-made processes. Persistent organochlorine compounds, e.g.

dioxin-like substances, and organochlorine pesticides, as well as heavy metals, e.g. mercury,

accumulate in the marine food chain. As a result, people consuming seafood are potentially

at increased risk to ingest simultaneously compounds that can have toxicological effects.

These conflicting facts form a potential base for an important public health conflict between

dietary recommendations on the one hand and toxicological safety assurance on the other

hand.

The introduction of this PhD-thesis starts with the position of omega-3 fatty acids in the

human diet, their role in the human body, and the current recommendations for omega-3

fatty acids. Next, an introduction to the nutritional-toxicological conflict related to seafood is

given, followed by a detailed description on how the evaluation of the intakes of nutrients

and contaminants will be performed in this PhD-thesis.


16

1. Omega-3 fatty acids and the human diet

Scientific evidence exists that long chain (LC) omega-3 poly-unsaturated fatty acids (PUFAs)

play an essential role in human health (Din et al, 2004; Kris-Etherton et al, 2002; Kris-Etherton

et al, 2003; Ruxton, 2004; Ruxton et al, 2004; Ruxton et al, 2005; Sidhu, 2003; Simopoulos et al,

2000). To situate these omega-3 PUFAs in the overall group of fatty acids, the table below

indicates the different types of fatty acids and gives some commonly occurring examples

(Table I.1).

Table I.1: The different types of fatty acids and some examples commonly occurring in food I. Saturated fatty acids II. Unsaturated fatty acids

Butyric acid C4:0 II.a. Mono-unsaturated fatty acids Caproic acid C6:0 Oleic acid C18:1n-9 Lauric acid C12:0 II.b. Poly-unsaturated fatty acids Myristic acid C14:0 II.b.1. Omega-6 fatty acids Palmitic acid C16:0 Linoleic acid C18:2n-6 Stearic acid C18:0 γ-linolenic acid C18:3n-6 Arachidonic acid C20:4n-6 II.b.2. Omega-3 fatty acids α-linolenic acid C18:3n-3 Eicosapentaenoic acid C20:5n-3 Docosahexaenoic acid C22:6n-3

The group of PUFAs is divided into two groups: omega-3 (n-3) and omega-6 (n-6) PUFAs,

differing in the position where the first double C-bound is located. Two PUFAs are called

‘essential fatty acids’ since they cannot be synthesized in the human body and are vital for

physiological integrity. Therefore, they must be obtained from the diet. One is linoleic acid

(LA, C18:2n-6) and belongs to the n-6 family. The other one is α-linolenic acid (LNA, C18:3n-

3) belonging to the n-3 family. These essential parent compounds can be converted in the

human body to LC fatty acids, but humans can not interconvert n-3 and n-6 fatty acids

(Ruxton et al, 2005). LA can be converted to arachidonic acid (AA, C20:4n-6) and further on

to longer chain derivates, and LNA to eicosapentaenoic acid (EPA, C20:5n-3) in a first step

and docosahexaenoic acid (C22:6n-3) in a next step. This conversion is summarized in Fig.

I.1.


17

Fig. I.1 Desaturation and elongation of n-6 and n-3 fatty acids (based on Din et al (2004))

Different enzymatic steps are needed to fulfil these conversions which include desaturation

and elongation. A crucial point is that the first enzymatic step in the desaturation of both LA

and LNA involves the enzyme Δ6 desaturase. This results in a competition of both substrates

for Δ6 desaturase. High intakes of LA have been suggested to decrease the desaturation of

LNA to EPA and DHA and favour higher levels of AA.

In developed countries dietary intakes of LA have increased over the last century from ~3%

to 5-7% of total energy intake due largely to an increased consumption of LA-rich vegetable

oils (Cunnane & Griffin, 2002; Liou et al, 2007; Ruxton et al, 2005). As a result, the conversion

of LNA to LC n-3 PUFAs in the human body is even more limited. In adult men conversion

to EPA is limited to approximately 8% and conversion to DHA is extremely low (< 0.1%). In

women conversion to DHA appears to be greater and affected by the physiological state (e.g.

higher conversion during pregnancy to meet the provision of foetal DHA) (Otto et al, 2001;

Williams & Burdge, 2006). It is questioned whether all human requirements for LC n-3

PUFAs can be met from the endogenous supply (Burdge & Wootton, 2002; Cunnane &

Griffin, 2002; Ruxton et al, 2005; Williams & Burdge, 2006). Therefore, most nutritional

guidelines now include recommendations for increased intakes of these fatty acids.

As already stated, over the past decades, the balance of fatty acids in the diet has shifted

away from n-3 PUFAs to n-6 PUFAs. Explanations for this shift include changing farming


18

practices resulting in meat being less rich in n-3 PUFAs, low consumption of seafood, and

developments in food technology leading to increased supply and high consumption of n-6

rich food products (e.g. margarines) (Ruxton, 2004; Simopoulos, 1999). As a result, the ratio

of n-6 to n-3 PUFAs increased in developed countries. Simopoulos (1999; 2002) indicates that

human beings evolved on a diet with a ratio of n-6 to n-3 fatty acids of around 1 to 4 whereas

in Western diets the ratio is now 15/1–16.7/1. Considering the dietary sources of n-3 PUFAs,

LNA can be found in plant sources, including green leafy vegetables, some seed oils (linseed

oil, rapeseed oil, flaxseed oil, and soybean oil) and nuts (e.g. walnuts). Synthesis of EPA and

DHA occurs in phytoplankton and animals, but not in plants. It is by planktivorous fishes

that LC n-3 PUFAs enter in the marine food chain and accumulate in seafood (mainly in oily

fish) (Cunnane & Griffin, 2002; Innis, 2007). Therefore, seafood is the food group with the

highest natural LC n-3 PUFA concentration, followed by poultry and eggs since poultry also

synthesise a considerable amount of LC n-3 PUFAs. Moreover, the fatty acids content of their

feed will influence their LC n-3 PUFA content.

2. The role of omega-3 fatty acids in the human body

Because of the changes in dietary and lifestyle patterns, chronic non-communicable diseases

- including obesity, diabetes mellitus, cardiovascular disease (CVD), hypertension and

stroke, and some types of cancer – have globally become increasingly significant causes of

disability and premature death (World Health Organization, 2003). In this context, the role of

seafood consumption and the relation of n-3 PUFAs with several diseases are summarized

below.

2.1. During development

Two major areas where n-3 PUFAs, particularly DHA, play a role are brain and eye function

(Ruxton, 2004). Therefore, the maintenance of an adequate level of DHA in both the brain

and the retina is important for proper nervous system and visual functions (Lands, 2005;

Salem & Pawlosky, 1992; Simopoulos, 1989). DHA accounts for 25% of the phospholipids in

cerebral grey matter, where it supports neural and brain activity (Ruxton et al, 2004; Ruxton

et al, 2005). A supply of n-3 PUFAs from seafood and other food items may have led to large


19

brain expansion during the long evolution of hominids to Homo sapiens. Studies suggest that

it is unlikely that the foetus can make sufficient DHA to support brain development. Given

that maternal DHA stores compensate for the limited ability of the foetus to synthesise DHA,

it is likely that an adequate intake of LC n-3 PUFA could impact foetal development (Otto et

al, 2001; Ruxton et al, 2005). However, the question on how important dietary DHA is during

human brain development remains unresolved (Innis, 2007).

2.2. Cardiovascular diseases

As mentioned before, the pioneering research of Bang and Dyerberg (1980) provided

evidence that the consumption of fish reduced the risk on heart disease in Greenland

Eskimos. The investigators proposed that this could be related to the high content of n-3

PUFAs in their diet. These findings initiated a lot of research to evaluate the effects of either

fish or n-3 PUFA consumption on human health. Several prospective epidemiological studies

have investigated the relationship between intake of n-3 PUFA and fatal coronary heart

disease (CHD). Nowadays a large body of evidence exists to suggest beneficial effects of fish,

oily fish and LC n-3 PUFAs on the molecular, cellular, and whole-body pathogenic processes

of atherosclerosis and thrombosis (Kris-Etherton et al, 2002; Kris-Etherton et al, 2003; Ruxton

et al, 2004; Williams & Burdge, 2006). Large-scale epidemiologic studies suggest that people

at risk for CHD benefit from consuming n-3 PUFAs from plant and marine sources (Kris-

Etherton et al, 2002). The ways in which n-3 PUFAs reduce CVD risk are still being studied.

Possible mechanisms of action of n-3 PUFAs in the prevention of cardiovascular disease are

(Din et al, 2004; Institute of Medicine, 2005; Kinsella et al, 1990; Kris-Etherton et al, 2002; Kris-

Etherton et al, 2003; Ruxton et al, 2005; Sidhu, 2003):

- antiarrhythmic mechanism: n-3 PUFAs decrease risk of arrhythmias that might be

implicated in sudden death;

- antithrombotic mechanism: n-3 PUFAs decrease risk of thrombosis, that might be

implicated in heart attack and stroke;

- antiatherosclerotic mechanism: n-3 PUFAs decrease the rate of growth of atherosclerotic

plaque;

- n-3 PUFAs lower triglyceride and remnant lipoprotein levels in a dose dependent

manner;

- n-3 PUFAs lower blood pressure;


20

- antiinflammatory mechanism: n-3 PUFAs reduce inflammatory responses, which is

important since inflammation has a central role in the development and progression of

CHD.

However, evidence from trials with n-3 PUFAs is less clear. Moreover, a recent meta-analysis

has revealed a lack of consistency between the different trials studying the influence of n-3

PUFAs on the risk of CHD (Hooper et al, 2006). The antiarrhythmic effect of fish oil remains

unproven although the idea is still viable and is being actively tested in further trials

(Brouwer et al, 2006). As a consequence, there is a need for further large-scale, well-executed,

randomised controlled trials to reach a consensus with regard to the role of these fatty-acids

in prevention of CHD and CVD in general (ISSFAL, 2007).

2.3. Other diseases

2.3.1. Inflammatory diseases

Lands (2005) stated that man must better balance n-3 and n-6 fatty acid intake to temper an

overactive eicosanoid system that leads to development of chronic inflammatory diseases.

Eicosanoids are produced from LC PUFAs and are biologically active substances which act

locally to influence a wide range of functions in cells and tissues, including inflammation.

There are different families of eicosanoids: prostaglandins, thromboxanes, leukotrienes, and

other metabolites. For each, there are separated series, derived either from a LC n-3 or LC n-6

PUFA. At sufficiently high intakes, LC n-3 PUFAs decrease the production of inflammatory

eicosanoids. Moreover, LC n-3 PUFAs give rise to a family of antiinflammatory mediators.

Modulation of LC PUFA intake, mostly by increasing the relative proportion of n-3 versus n-

6 PUFAs, may as such play a role in the prevention of diseases as asthma, rheumatoid

arthritis, and bowel disease (Calder, 2006; De Caterina & Basta, 2001). Supplementation with

large doses of fish oil has been shown to relieve symptoms in patients with rheumatoid

arthritis but findings have been far less consistent for other inflammatory diseases (de

Deckere et al, 1998). In the future, more, better designed and larger clinical trials are needed

to assess the therapeutic potential of LC n-3 PUFAs in inflammatory diseases (Calder, 2006).


21

2.3.2. Cancer

Increasing evidence from animal and in vitro studies indicates that n-3 PUFAs present in

fatty fish and fish oil inhibit carcinogenesis (Larsson et al, 2004; Terry et al, 2003). Ecologic

studies have shown that high per capita fish consumption is correlated with a lower

incidence of cancer in the populations (Caygill et al, 1996; Hursting et al, 1990; Kaizer et al,

1989; Sasaki et al, 1993). Several mechanisms whereby n-3 PUFAs may modify the

carcinogenic process have been proposed. These include suppression of AA-derived

eicosanoid biosynthesis, influences on transcription pathways, alteration of oestrogen

metabolism, increased or decreased production of free radicals and reactive oxygen species,

and mechanisms involving insulin sensitivity and membrane fluidity. On the basis of these

multiple mechanisms, n-3 PUFAs may have an important influence on carcinogenesis.

Further studies are needed to identify new mechanisms and to evaluate and verify the

mechanisms in humans to gain more understanding of the effects of marine n-3 PUFA intake

on cancer risk for humans (Larsson et al, 2004; Terry et al, 2003).

2.3.3. Mental function and mood

Research related lower intakes of LC n-3 PUFAs with mental illness. Epidemiological studies

have found that countries with high seafood consumption tend to have a low prevalence of

major depression (Hibbeln, 1998). However, Appleton et al (2006) reported in a recent review

that trial evidence that examines the effects of n-3 PUFAs on depressed mood is limited and

is difficult to evaluate because of considerable heterogeneity. The evidence available

provides little support for the use of n-3 PUFAs to improve depressed mood. Furthermore,

indications exist that imbalances in PUFA status could be linked to behavioural and learning

disorders such as attention deficit hyperactivity disorder, dyslexia, dyspraxia, and autism

(Richardson, 2004a; Richardson, 2004b; Richardson & Montgomery, 2005). The influence of

higher intakes of LC n-3 PUFA is not clear yet. Also for dementia it is claimed that increasing

prevalence is linked with low oily fish consumption (Ruxton, 2004; Ruxton et al, 2005).

Nevertheless, there is a need for more evidence from intervention studies, particularly

relating to postnatal depression, child intelligence, and ageing where LC n-3 PUFAs could

have far-reaching benefits (Ruxton et al, 2005).


22

3. Recommendations for omega-3 fatty acids

Due to the positive health effects of n-3 PUFAs, recommendations for dietary intake have

been formulated. In the past, dietary allowances have been developed with the aim of

preventing signs or symptoms of deficiency. However, in the new guidelines the American

Institute of Medicine (IOM) requires that ‘reduction in the risk of chronic degenerative

disease must be included in the formulation of dietary recommendations’ (Gebauer et al,

2006; Institute of Medicine, 2005), being very relevant for the recommendations related to

fatty acids. It is worthwhile to review some of the existing, international n-3 and other PUFA

recommendations.

(1) The International Society for the Study of Fatty Acids and Lipids (ISSFAL) gives the

following specific recommendations about the intake of PUFAs for healthy adults:

1. An adequate intake of LA equal to 2% of the total energy intake (%E);

2. A healthy intake of LNA equal to 0.7 %E;

3. For cardiovascular health, a minimum intake of EPA and DHA combined: 500 mg/d.

Moreover, the ISSFAL recognises that there may be a healthy upper limit to the intake of LA

but insufficient data exist at present to set a precise value on such an upper limit (ISSFAL,

2007).

(2) The American IOM used the median PUFA intakes of the American population since n-6

and n-3 deficiency is nonexistent in healthy individuals. In summary, they determined an

acceptable macronutrient distribution range (AMDR) of 5 to 10 %E for n-6 PUFA and 0.6 to

1.2 %E for n-3 PUFA. In both cases they stated that approximately 10% of the AMDR can be

consumed as LC PUFAs (Institute of Medicine, 2005).

(3) In the UK a recommendation of 0.2 g LC n-3 PUFA per day was used until recently.

However, new advice to increase fatty fish consumption led to new recommendations,

suggesting that one to four portions of fatty fish per week should be consumed by the

general UK population. This should be equal to an intake of 450 mg LC n-3 PUFA per day

(SACN/COT, 2004).

(4) The World Health Organization (WHO) recommended an intake of n-6 PUFAs and n-3

PUFAs equal to 5 to 8%E and 1 to 2%E, respectively (World Health Organization, 2003).


23

This illustrates that the international recommendations on dietary PUFA intake are quite

diverse in quantity as well as mode of expression. In this PhD-thesis, the recommendations

formulated by the Belgian Health Council are applied. In a nutshell, for adults a dietary

reference intake (DRI) of 4 to 8 %E is recommended for n-6 PUFA, with a minimum of 2 %E

coming from LA and a DRI of 1.3 to 2.0 %E is recommended for n-3 PUFA with a minimum

recommended intake of 1 and 0.3 %E for LNA and EPA+DHA, respectively (Belgian Health

Council, 2007).

4. Seafood, the nutritional-toxicological conflict

To avoid confusion about terminology the generic term ‘seafood’ will be used in this PhD-

thesis to describe the group of fish and shellfish, the first being the sum of marine fishes and

fresh water fishes and the latter being the sum of molluscs, crustacean, and cephalopods.

Sometimes, the term ‘marine foods’ is used as an alternative for seafood. However, in

literature the term ‘fish’ is mostly used when describing the relation of seafood with health

and/or diseases. Reason for this is that the beneficial effects are related to EPA and DHA

which are most abundant in fatty fish species. Seafood species other than fish (e.g. crustacean

and molluscs) have in general a lower fat and n-3 PUFA content and are in general also less

consumed.

There is no doubt that seafood is the most important source of LC n-3 PUFAs. Together with

the fact that there is an imbalance between the current intake of LC n-3 PUFA in developed

countries and the recommendations, and that an increase of the n-6/n3 PUFA ratio in most

Western countries is seen, a further stimulation in seafood consumption should be an

evident solution. In addition, seafood is a good source of high-quality proteins and

micronutrients, e.g. iodine, vitamin D, selenium, and zinc.

However, during the last decades an increasing number of data on high levels of

contaminants in seafood are published in scientific literature as well as in public media

(Ashizuka et al, 2005; Corsolini et al, 2005; Foran et al, 2004; Hites et al, 2004a; Hites et al,

2004b; Isosaari et al, 2004; Karl et al, 2002; Sidhu, 2003; Simm et al, 2006). Due to the industrial

revolution, oceans, seas, lakes, and rivers are contaminated with many persistent, chemical

contaminants. Due to biomagnification and bioaccumulation, these contaminants are


24

concentrated in the aquatic food chain to concentrations that could form a health risk for

consumers (Burreau et al, 2006). As a result, increased seafood consumption to achieve an

adequate LC n-3 PUFA intake may simultaneously increase the contaminant intake to levels

of toxicological concern. On the other hand, consumers decreasing their seafood intake in

order to avoid contaminant exposure may be incurring an inadequate intake of LC n-3

PUFAs (Cohen et al, 2005). This nutritional-toxicological conflict forms the base of the study

elaborated in this PhD-thesis and is summarized in the scheme below (Fig. I.2).

Fig. I.2 Nutritional-toxicological conflict linked to increased seafood consumption

Due to this nutritional-toxicological conflict, it was of interest to evaluate the risks and

benefits related to the consumption of seafood in order to formulate safe recommendations

about seafood consumption for the Belgian population. Also on international level, the FAO

and WHO have demanded consultation of an expert group to study the risks and benefits

related to the consumption of seafood. The latter is already discussed at the Commission of

the Codex on Contaminants in Food (Commission of the Codex Alimentarius). Planning of

the work is ongoing and some internal working groups at the level of the Food and

Agriculture Organization (FAO) and WHO are now established, showing the pertinence of

this subject.

Seafood consumption

LC n-3 PUFAs

Contaminants (dioxins, mercury, …)

Negative to human health

Positive to human health

Nutritional-toxicological conflict


25

5. Evaluation of risks and benefits

In this PhD-thesis the focus lies on the assessment of exposure to compounds ingested by

food. Other ways of exposure are not considered. Food contains a wide range of substances

which are either desired (nutrients, additives, and other components) or undesired, such as

natural toxins, pesticide residues, mycotoxins, or contaminants. For all these materials,

excessively high but in some cases also insufficiently low amounts can create a risk (Kroes et

al, 2002). In the EU Regulation (EC) No 178/2002, risk is defined as ‘a function of the

probability of an adverse health effect and the severity of that effect, consequential to a

hazard’ (European Commission, 2002). Regulatory measures taken solely from a risk point of

view could restrict the consumption of a given food, whereas the health consequences of not

eating that food might be more serious than the risk. Therefore, it is important to include also

the benefits in such an evaluation.

Risk assessment is defined as a four-step process:

Hazardidentification

Hazardcharacterisation

Exposureassessment

Riskcharacterisation

11

22

33

44

The latter step, risk characterisation, integrates the information collected in the preceding

three steps (European Commission, 2002). It interprets the qualitative and quantitative

information on the toxicological properties of a chemical in combination with the assessed

amount to which individuals (parts of the population or the population at large) are exposed

to this chemical (Kroes et al, 2002). On a colloquium of the European Food Safety Authority

(EFSA) held in July 2006, experts and scientists came to a general consensus that a risk-

benefit evaluation should mirror the paradigm already well established for risk analysis.

This means that the benefit assessment part of the risk-benefit assessment should include

benefit identification, benefit characterisation (dose-response assessment), exposure

assessment, and (probability for) benefit characterisation. An introduction to the four

different steps of the risk-benefit evaluation applied to the theme ‘risks and benefits related

to seafood consumption’ is given below.


26

5.1. Hazard and benefit identification

The hazard and benefit identification concerns the determination of the

different compounds to include in the analysis of the risks and benefits related

to seafood.

5.1.1. Nutrients – benefit identification

Several nutritional benefits are related to seafood. Seafood contains very high quality

proteins, being second only to egg protein in digestibility and supporting growth. With

respect to vitamin content, seafood is an excellent source of the B vitamins niacin and B12,

and is in general a better source of vitamins D and A compared to other protein sources as

beef, pork, or chicken. Seafood can also contribute appreciable amounts of heme iron and

zinc, nutrients that tend to be low in people's diets and it is among the best sources of dietary

selenium.

In this PhD-study three nutritional compounds were studied in detail, chosen since they are

present in seafood in a relatively high concentration compared to other food items. The first

is the sum of EPA and DHA, the LC n-3 PUFAs most abundantly present in seafood.

Furthermore, vitamin D and iodine were taken into account as well. The benefit

identification of EPA and DHA is already described in previous parts of this introduction.

Therefore, only vitamin D and iodine are considered below.

5.1.1.1. Vitamin D

Seafood - especially oily fish - is frequently regarded as the most important food source of

the fat soluble vitamin D. The letter D origins from the German word ‘Dörschleberöl’, which

means cod liver oil, hereby showing the historical link between vitamin D and seafood.

Vitamin D is not an essential vitamin sensu stricto, since ultraviolet (UV) induced skin

production constitutes the main contributor to vitamin D in humans who are sufficiently

exposed to sunlight (Brustad et al, 2004). Apart from this endogen synthesis, food also

provides vitamin D. Nevertheless, next to oily fish only a relatively small number of food

items such as eggs, liver, and butter contain nutritionally significant quantities of vitamin D


27

(Bender, 2002; Lamberg-Allardt, 2006; Suzuki et al, 1988). To avoid vitamin D deficiency in

Belgium all margarines and minarines are currently fortified with vitamin D (6,25 - 7,5 µg

vitamin D per 100 g margarine) (Warenwetgeving, 1998).

The principal physiologic function of vitamin D is to maintain calcium homeostasis. Vitamin

D functions as a hormone, synthesized far from the sites of biological action and reaching

these distant sites through the blood stream. Vitamin D acts by binding to vitamin D

receptors principally located in the nuclei of target cells. Due to that binding, the receptor

will act as a transcription factor that modulates the gene expression of transport proteins, e.g.

calbindin (Wolpowitz & Gilchrest, 2006). Severe deficiency of vitamin D leads to rickets in

children and osteomalacia in adults (Bender, 2002). Yet, an adequate dietary intake of

vitamin D is of great importance for young children, pregnant women, and elderly people.

Besides its role in the calcium homeostasis, more extensive roles for vitamin D were

suggested by the discovery of the vitamin D receptor in tissues that are not involved in the

calcium metabolism. As such, the role of vitamin D in regulation of the immune system and

its possible role in the prevention and treatment of cancer and immune-mediated diseases

was discovered (Mullin & Dobs, 2007).

5.1.1.2. Iodine

Oceans are considered as a huge natural reservoir of iodine. Iodine is distributed from the

ocean in the atmosphere by evaporation and by rain it returns to earth. The biological

function of iodine in the human body relates to its incorporation in the thyroid hormones

(Institute of Medicine, 2005). The iodine content in most foods is low and can be affected by

the content of the soil, irrigation, and fertilizers. Foods of marine origin have higher

concentrations of iodine because marine species concentrate iodine from the seawater (Dahl

et al, 2004; Karl et al, 2001). Another important dietary source is dairy food, due to the

secretion of iodine in cow’s milk. Lack of sufficient iodine results in inadequate thyroid

production leading to the so-called iodine deficiency disorders, including mental retardation,

hypothyroidism, goitre, cretinism, and varying degrees of other growth and developmental

abnormalities (Institute of Medicine, 2005).


28

5.1.2. Contaminants – hazard identification

At the contaminant side, two different categories of persistent, chemical contaminants were

considered: non-carcinogenic (e.g. mercury) and carcinogenic (e.g. dioxins) compounds.

Although a very wide range of contaminants may accumulate in the marine food chain, it

was decided to limit this study to the relevant ones for which enough data were available in

the international scientific literature and other important publicly available data sources.

5.1.2.1. Mercury

The rationale to include mercury (Hg) is that seafood is the most important source of Hg in

the human food chain. Moreover, in the marine environment, inorganic Hg is to high extent

transformed to methyl mercury (MeHg), which further accumulates in the marine food chain

(European Commission, 2006). This organic MeHg is very toxic for humans (Clarkson &

Magos, 2006; EFSA, 2004). The primary target of MeHg is the central nervous system and the

developing brain is thought to be the most sensitive target organ for MeHg toxicity (EFSA,

2004). In addition, there are indications that Hg can inhibit the preventive role of LC n-3

PUFAs on cardiovascular diseases (Chan & Egeland, 2004; Salonen et al, 2000).

5.1.2.2. PCDFs, PCDDs, PCBs

Next to Hg, two groups of persistent organic pollutants are considered in this work, being

the polychlorinated dibenzo-p-dioxins (PCDDs) and polychlorinated dibenzofurans (PCDFs)

on the one hand and the polychlorinated biphenyls (PCBs) on the other hand. PCDDs,

PCDFs and PCBs are ubiquitous in the environment and occur at picomolar levels in foods.

PCDDs and PCDFs are by-products of combustion and of various industrial processes and

thus unintentionally present in the environment. In contrast, PCBs were manufactured in the

past for a variety of industrial uses, notably as electrical insulators or dielectric fluids and

specialized hydraulic fluids. Most countries banned manufacture and use of PCBs in the

1970s. However, past improper handling of PCBs constitutes a continuing source of these

compounds in the environment. Exposure of a typical person to these persistent organic

pollutants occurs primarily through foods (> 90%), particularly animal fats because of the

lipophilic characteristics of these substances. Bioaccumulation of dioxin-like compounds

through the food chain begins at the point of consumption of contaminated plants, soil, or

sediment by animals, which then are used to produce food for humans (EFSA, 2005c; World

Health Organization, 2007; Yaktine et al, 2006).


29

The reason to consider these substances in this case study is that seafood has a higher

concentration of these compounds per gram fat compared to other food items. Last decennia,

a decreasing trend of the levels of PCBs, PCDDs, and PCDFs in food items is found, mainly

due to strict rules and regulations to reduce human intake of these compounds. However,

the decreasing trend seems to be quite limited in the aquatic environment since these

compounds seem rather stable in large reservoirs like seas and rivers (AFSSA, 2005). As a

result, seafood is currently one of the most important contributors to the total dietary intake

of PCBs, PCDDs, and PCDFs (AFSSA, 2005; Bilau et al, 2007; Bocio & Domingo, 2005;

Darnerud et al, 2006; Fattore et al, 2006; Kiviranta et al, 2004).

There exist 75, 135, and 209 different congeners of PCDDs, PCDFs, and PCBs, respectively,

depending on the number and position of the chlorine atoms (Fig. I.3).

O

O19

46

8

7

2

3

QsQv O

19

46

8

7 3

2

QsQv ortho

para

meta

14

65

23orthometa

ortho

para

meta

1'4'

6' 5'

2' 3'ortho meta

QsQv PCDDs PCDFs PCBs

Fig. I.3 Chemical structures of polychlorinated dibenzo-p-dioxins (PCDDs), polychlorinated dibenzofurans

(PCDFs), and polychlorinated biphenyls (PCBs); Qs and Qv represent chlorine substitution

Only 7, 10, and 12 of the existing PCDD, PCDF, and PCB congeners, respectively, exhibit

relevant dioxin-like activity. The subset of 7 PCDDs and 10 PCDFs correspond to chlorine

substitution at the 2, 3, 7, and 8 position (hereafter referred to as PCDD/Fs). The 12 dioxin-

like PCBs (dl PCBs) included have either one or no chlorine substitution in the ortho position

(non-ortho or mono-ortho PCBs) (World Health Organization, 2007).

For these 29 dioxin-like compounds, toxic equivalency factors (TEFs) for mammals have

been derived (Van den Berg et al, 1998; Van den Berg et al, 2000; Van den Berg et al, 2006).

The TEF approach relates the toxicity of these chemicals to that of 2,3,7,8-tetrachlorinated

dibenzo-p-dioxin. In the TEF concept, the assumption is made that PCDD/Fs and dl PCBs

have a common mechanism of action, which involves binding on the aryl hydrocarbon

receptor, an intracellular receptor protein. This binding is considered to be the necessary first

step in expressing the toxicity of these compounds as well as for the assumption that the


30

toxic effect of the different congeners is additive. Many uncertainties exist in the use of the

TEF approach for human risk assessment, but pragmatically it is the most feasible approach

(World Health Organization, 2007). TEFs allow determination of the toxic equivalent (TEQ)

concentration of the residue. For a particular residue containing a mixture of i PCDDs, j

PCDFs, and k PCBs, the TEQ is calculated according to the following equation (World Health

Organization, 2007):

∑∑ ∑ ++=k

kkji j

jii TEFPCBTEFPCDFTEFPCDDTEQ )*()*()*(

with TEQ the toxic equivalent (pg/g) of a mixture of i PCDDs (pg/g), j PCDFs (pg/g), and k

PCBs (pg/g) all with their respective TEF value (dimensionless). Exposure to a mixture of

PCDD/Fs and dl PCBs causes dermal toxicity, immunotoxicity, carcinogenicity,

reproductive and developmental toxicity, and disruption of endocrine functions (World

Health Organization, 2007).

Besides the twelve dl PCB congeners, the non dioxin-like (ndl) PCBs were considered as

well. Ndl PCBs also cause multiorgan toxicity, the mechanism of which is not known. Much

higher doses are needed than for the dl PCBs, but only very few data on a few individual

congeners are available. Much more work has to be done before a similar quantitative risk

assessment as for dioxin-like substances can be performed. Data on the occurrence of ndl

PCBs in food and feed are often referred to as the sum of six or the sum of seven indicator

PCBs (Σ6 iPCBs = congeners 28, 52, 101, 138, 153, and 180 or Σ7 iPCBs= congener 118 + Σ6

iPCBs, with congener 118 being a dl PCB). In this PhD-thesis, the Σ7 iPCBs were considered.

In general, Σ6 iPCBs represent at least about 50% of the total amount of ndl PCBs in food

(EFSA, 2005c).

5.2. Hazard and benefit characterisation

Hazard characterisation involves the quantitative description of the level at

which a compound has potential to cause adverse effects (based on dose-

response relationships). Likewise, benefit characterisation can be defined as the

quantitative description of the level at which a nutrient intake is sufficient to

meet the requirement of all healthy individuals to stay in good health and to

avoid deficiency. Details about dose-response relationships are beyond the


31

scope of this thesis and are therefore not described. However, two types of quantitative

levels are applied in this PhD-thesis to evaluate nutrient and contaminant intakes related to

seafood consumption; first, the DRIs for nutrients and second, the tolerable daily or weekly

intakes (TDI or TWI) for contaminants. These reference levels are described in more detail in

chapter II and IV.

5.3. Exposure assessment

The largest part of the work executed in this PhD-thesis is related to the

exposure assessment. The latter has been defined by the WHO as the

qualitative and/or quantitative evaluation of the likely intake of biological,

chemical or physical agents via food as well as exposure from other sources if

relevant (World Health Organization, 1997). In this PhD-thesis, only the

exposure via food – and for the risk-benefit analysis only via seafood – is taken

into consideration. The term ‘exposure assessment’ is often replaced by ‘intake assessment’

when considering only food as exposure route.

Dietary intake of nutrients and contaminants can be estimated in different ways. The most

direct estimate is by measuring concentrations in duplicate diets collected over a certain

time. This is a technique in which equivalent portions of all foods and beverages consumed

by an individual are collected for direct analysis in order to estimate the individual’s intake

of energy, nutrients, and/or other food components (Cameron & Van Staveren, 1988).

However, dietary intake is usually estimated by models combining data on food

consumption with concentration data measured in foods and food groups (World Health

Organization, 2007). The latter method was applied in this thesis and involved three

important steps:

1. the collection of food consumption data;

2. the collection of nutrient and contaminant data; and

3. an appropriate method for combining both sources of data and assessing the intake.


32

5.3.1. Consumption data

In principle, to assess food consumption four different types of data can be used:

1. food supply data;

2. data from household consumption surveys;

3. data from dietary surveys of individuals;

4. the collection of duplicate diets (Kroes et al, 2002).

For the case study investigated in this PhD-thesis, data from dietary surveys were most

appropriate. Different methods can be applied to assess such data, to be divided in two

categories: record and recall methods.

- Record data collect information on current consumption over one or more days.

- Recall methods reflect past consumption, varying from intake over the previous day (24-

hour recall) to usual food intake (food frequency questionnaire or dietary history) (Kroes

et al, 2002; Willett, 1998).

Four different methods within the category of dietary surveys are distinguished:

(1) Food records (also called dietary records or food diaries) are kept for a specified time

period, usually 1 to 7 days. If total daily intake is required, the food records should include

all foods and beverages consumed at meals and in between, in quantified amount (Kroes et

al, 2002; Willett, 1998).

(2) In the 24-hour recall method the subject is asked by a trained interviewer to recall and

describe the kinds and amounts of all foods and beverages ingested during the immediate

past, mostly a 24- or 48-hour period. Food quantities are usually assessed by using

household measures, food models, or photographs (Kroes et al, 2002; Willett, 1998).

(3) A food frequency questionnaire (FFQ) consists of a structured list of individual foods or

food groups. The aim of the FFQ is to assess the frequency with which these items are

consumed during a specified period (e.g. daily, weekly, monthly, yearly). FFQs may be

qualitative, semi-quantitative or completely quantitative. Qualitative FFQs only obtain the

usual number of times each food is eaten during a specified period. Semi-quantitative

methods allow estimation of a standard portion or ask respondents to indicate how often

they consume a specified common amount. A quantitative FFQ allows the respondent to

indicate any amount of food typically consumed. The FFQ is often used to rank individuals


33

by food or nutrient intakes and also by food group intakes so that high and low intakes may

be studied (Kroes et al, 2002; Willett, 1998).

(4) With the aid of a dietary history method, a trained interviewer assesses an individual’s

total usual food intake and meal pattern. The respondent is asked to provide information

about his/her pattern of eating over an extended period of time (often a typical week) and

also to recall the actual foods eaten during the preceding 24 hours. In addition, the

interviewer completes a checklist of foods usually consumed. Finally as a cross-check, the

respondent is often asked to complete a three-day estimated record (Kroes et al, 2002; Willett,

1998).

It should be noted that there is no single ideal method to assess food consumption. The

choice depends on the objectives of the study, the foods of primary interest, the need for

group versus individual data, the characteristics of the population, the time frame of interest,

the level of specificity needed for describing foods, and available resources. Currently, most

methods to assess food consumption are not developed explicitly from the perspective of

risk assessment and the available data are used for other purposes than the original ones as

well (Kroes et al, 2002).

5.3.2. Nutrient and contaminant concentration data

A key component in nutrient and contaminant intake assessment is the collection of

quantitative data being accurate (i.e. agreeing with the actual concentration) and

representative (i.e. reflecting the concentration of the whole group) whether it be nutrients or

contaminants. In order to collect data, two different strategies can be applied. First, nutrient

and contaminant concentrations can be measured in representative samples. Such an

approach is scientifically preferable, but very expensive and laborious. Samples can be

collected on the base of a representative sampling plan or by the application of a duplicate

diet approach or a total diet study. The latter methodology is recommended by the FAO and

WHO for estimating dietary exposure of the population. In total diet studies, representative

samples of widely consumed foods are collected and analyzed for the constituents of interest

(Kroes et al, 2002). Alternatively, data found in secondary sources like food composition

databases (in the case of nutrients) and scientific literature, existing databases, and published

reports (in the case of contaminants) can be used.


34

For the intake assessment executed in this thesis, the second strategy of using secondary data

was adopted. This is in accordance with the strategy recently proposed by Brüders et al

(2005), stating that existing data should be used in the most effective way as collecting

samples and analyzing them is expensive. Nevertheless, the comparison of data from

different sources is always difficult especially if one only has access to aggregated data

without having sufficient information about the collecting scheme, location data, sampling

procedure, or laboratory procedures (Brüders et al, 2005). Detailed information about the

data collection and data handling as applied in this PhD-thesis is reported in chapter III.

5.5.3. Methodologies for intake assessment

In its simplest form, the model to represent dietary intake/exposure can be considered as:

Consumption x (Concentration or Residue) = Dietary exposure.

There are, however, three different approaches for combining the consumption data with

nutrient/contaminant concentrations: deterministic modelling, simple distribution, and

probabilistic modelling. Both the simple distribution and probabilistic approach are applied

in this PhD-thesis.

5.5.3.1. Deterministic modelling

Deterministic modelling involves using a point estimate of each variable within the model

(Vose, 1996). In the context of intake assessment, the term ‘deterministic modelling’ refers to

a method whereby a fixed value for food consumption (such as the average or high level

consumption value) is multiplied by a fixed value for the concentration and the intakes of all

sources are then summed. In a schematic way, it can be summarized as follows:

μ(X) x μ(C) = μ(Y) Mean scenario

Max(X) x Max(C) = Max(Y) Worst case scenario

With X= consumption of a certain food item or food group, C= concentration of the

considered compound in that food item, Y= exposure to the considered compound via the

considered food item or food group.


35

Deterministic modelling is commonly used as a first step in exposure assessment because it

is relatively simple and inexpensive to carry out. However, this approach does not provide

insight into the range of possible exposures that may occur in a population (Kroes et al, 2002;

Lambe, 2002) and it also obscures the ability to determine which scenarios present a risk that

is likely to occur (Petersen, 2000).

5.5.3.2. Simple distribution

‘Simple distribution’ is a term used to describe a method that employs distributions of food

intake but uses a fixed value for the nutrient and contaminant concentration variables (Fig.

I.4).

Fig. I.4 Scheme of the simple distribution approach

The results are more informative than those of the deterministic approach because they take

into account the variability that exists in food consumption patterns (Kroes et al, 2002;

Lambe, 2002). This approach is widely used for the intake assessment of macro- and

micronutrients, where one value is used to represent the content of each nutrient in each

food. This methodology was applied in this thesis to assess the intake of PUFAs via the total

diet for different subgroups of the Flemish population and to assess the vitamin D intake of

Flemish adolescents (chapter II).

5.5.3.3. Probabilistic modelling

In contrast to the deterministic approach, probabilistic modelling involves incorporation of

the variability and/or uncertainty for the different parameters. As such, it takes into account

all the possible values that each variable could take and weights each possible model

outcome by the probability of its occurrence (Vose, 1996) (Fig. I.5).

Fig. I.5 Scheme of the probabilistic approach


36

Food consumption and nutrient/contaminant concentration data can be entered in

probabilistic models by two different approaches: non-parametric or parametric. The choice

of the approach depends largely on the data resources available.

- In the non-parametric approach all the individual data are used as such, without

assuming any underlying probability model.

- In the parametric approach probability distributions are fitted to the available data. This

process involves making assumptions about the underlying mathematics of the

distribution. In a next step, random values are drawn from these distributions and used

as input for the mathematical model which describes the intake assessment process

(Lambe, 2002).

The probabilistic analysis for intake assessment, requiring appropriate modelling software,

permits the exposure assessor to consider the whole distribution of exposure, from minimum

to maximum, with all modes and percentiles (Kroes et al, 2002). A probabilistic approach was

applied in this thesis to assess the intake of multiple compounds of interest via seafood

consumption (chapter IV). In this PhD-thesis, only variability (not uncertainty) was taken

into account in the probabilistic intake assessment.

5.4. Characterisation of risks and benefits

Risk characterisation is the process of comparing the modelled estimates to

toxicological endpoints and establishing the relevance for human health. In this

PhD-thesis, a benefit characterisation was elaborated by comparing the

assessed nutrient intakes with nutritional recommended endpoints. The results

of this evaluation are described in chapter IV.


37

6. The public health relevance of this work

In very general terms, public health is the science that deals with optimisation of the health

and wellbeing in the community at large, operating through organised societal efforts and

diverse structural measures, including regulatory initiatives, health promotion campaigns,

educational efforts, etc. Within this broad scientific area, both nutritional and environmental

issues are important determinants and therefore deserve continued attention from scientific

and from policy perspective.

A very important aspect within the broad instrumentarium of public health research is the

assessment of ‘exposure’ in the general population and the development of algorithms that

allow for the interpretation of such exposures in a broader context of potential health risk.

This PhD-thesis is situated at the crossroads of nutritional and environmental epidemiology

as it merges know-how from both areas and combines them in an attempt to optimise the

understanding of their impact on public health. For example, some important topics that

were handled within this PhD-thesis are:

- data management work leading to extensive databases relevant in exposure assessment

(see chapter III),

- set up of a methodology and software module for probabilistic exposure assessment

(chapter IV), and

- scenario analyses to evaluate the impact of certain dietary interventions (chapter IV).

The ‘public health policy cycle’ is an important concept that has been designed as a helpful

tool when developing a strategy to tackle a public health problem. This cycle can – somewhat

artificially - be broken down into seven different steps (Margetts, 2004):

1. Identify key nutrition-related problem

2. Set goal

3. Define objectives for goal

4. Create quantitative targets

5. Develop program

6. Implement program

7. Evaluate program

The work in this thesis is predominantly focusing on steps 1 and 7 and can also be of help in

the development of interventional strategies.

Caraher & Coveney (2004) argue moreover that public health nutrition, through the medium

of health promotion, needs to address the wider issues of who controls the food supply and


38

thus the influences on the food chain and the food choices of the individual and

communities. Therefore, food policy should seek to make the social infrastructure conductive

to health decisions about food (Caraher & Coveney, 2004). In the general discussion of this

PhD-thesis (chapter V) some of these wider issues are discussed.

7. Outline of the thesis

Chapter II describes the intake of PUFAs via the total diet assessed for different subgroups

of the Flemish population: Flemish preschool children, adolescents from the Ghent region,

and young women from the Ghent region. Some thoughts relevant to the whole Belgian

population are added in the discussion part of this chapter. Chapter II gives an indication on

the contribution of seafood to the total intake of LC n-3 PUFAs and the overall adequacy of

the PUFA intake compared to the recommendations. For the adolescent population, the

importance of seafood as vitamin D source is described as well.

It is important to mention that chapter II of the thesis differs from the other parts in the way

that for the analyses described in this chapter the total diet was taken into account (so no

focus on seafood only). Furthermore, the intake was measured with the simple distribution

approach whereas further on in the thesis a probabilistic approach was used to calculate

nutrient and contaminant intakes via seafood consumption only.

Chapter III focuses on different important methodological aspects encountered during the

setup of the different databases needed for the probabilistic intake assessment of nutrients

and contaminants via seafood only. It describes the process of collecting data on the origin,

the nutrient concentration, and the contaminant concentration of seafood available on the

Belgian market.

Chapter IV describes the results of the probabilistic intake assessment of nutrients and

contaminants via seafood consumption only. First, the current intake based on existing food

consumption data for the Belgian population was described, followed by an evaluation of

the benefits and risks related to the assessed intakes. Second, different consumption

scenarios were considered in order to find out whether it is possible to meet the current


39

recommendation for LC n-3 PUFA intake by the consumption of seafood only, without

increasing the intake of contaminants to levels of toxicological concern.

Important discussion points related to the subject of this thesis (i.e. seafood consumption and

intake of n-3 PUFAs) but not investigated in detail in the discussion parts of the individual

chapters are described in chapter V, for instance the intake of dioxin-like compounds via the

total diet, consumers’ perception related to seafood, the importance of fortified foods and

supplements, and sustainable aspects related to fisheries.

Chapter II. PUFA and vitamin D intake and importance of seafood as nutrient source

41

Chapter II. Overall dietary PUFA intake and importance of seafood as PUFA and vitamin D source

This chapter is based on the following papers:

1. Sioen I, Pynaert I, Matthys C, De Backer G, Van Camp J, De Henauw S. Dietary intakes

and food sources of fatty acids for Belgian women, focused on n-6 and n-3

polyunsaturated fatty acids. Lipids 2006; 41(5): 415-422.

2. Sioen I, Matthys C, De Backer G, Van Camp J, De Henauw S. Importance of seafood as

nutrient source in the diet of Belgian adolescents. Journal of Human Nutrition and Dietetics

2007; In Press.

3. Sioen I, Huybrechts I, Verbeke W, Van Camp J, De Henauw S. Omega-6 and omega-3

poly-unsaturated fatty acid intakes of pre-school children in Flanders-Belgium. British

Journal of Nutrition 2007; 98(4): 819-825.


42

1. Introduction

Despite the potential favourable health effects, it has been repeatedly reported that modern

diets in developed countries are low in omega-3 poly-unsaturated fatty acids (PUFAs)

(Muskiet et al, 2006; Simopoulos, 2002). The diet in developed countries has evolved from a

diet rich in α-linolenic acid (LNA) and long chain (LC) PUFAs to a modern Western diet in

which linoleic acid (LA) is the main PUFA (Ruxton et al, 2004; Sontrop & Campbell, 2006).

This is mainly due to intensified agricultural production processes and the increased

consumption of margarines and food items containing vegetable oils that are rich in LA. A

result of this dietary evolution is an increased ratio of omega-6 (n-6) over omega-3 (n-3)

PUFAs, which has been suspected to promote the pathogenesis of many diseases, including

cardiovascular and inflammatory diseases (Simopoulos, 2002).

In order to develop appropriate interventional strategies at the population level, it is

important first to quantify the differences between current dietary intakes and

recommendations (American Heart Association et al, 2006). Therefore, it was of interest to

accurately estimate the current intakes of individual PUFAs of different subgroups of the

Belgian population. Moreover, this PhD-thesis assesses quantitatively the seafood

contribution to the total intake of LC n-3 PUFAs and vitamin D, based on available food

consumption data.

The investigations described in this chapter focus on:

1. The dietary PUFA intake of Flemish pre-school children, the food sources that contribute

to the intake of the individual n-6 and n-3 PUFAs as well as potential ways to bridge the

gaps between current intakes and recommended levels. To the authors’ knowledge, it was

the first time that individual PUFA dietary intakes of pre-school children in Europe were

described. In the past, only individual PUFA intake data of Australian, Canadian, and

Chinese children have been published (Barbarich et al, 2006; Innis et al, 2004; Meyer et al,

2003).

2. The importance of seafood consumption as nutrient source in the diet of Flemish

adolescents. The intake of individual n-6 and n-3 PUFAs as well as vitamin D was

investigated in detail, the food sources of the considered nutrients were studied and it was


43

explored whether increased seafood consumption can be a solution for low intake of LC n-3

PUFAs and vitamin D.

3. The total intake of n-6 and n-3 PUFAs and the main food sources of these fatty acids in a

population sample of Flemish young women, i.e. women of childbearing age.

For this investigation, available Flemish data were used. The results were enriched with data

of the Belgian Health Interview Survey (Scientific Institute of Public Health, 2006) and the

Belgian Food Consumption Survey (De Vriese et al, 2006) in order to generalise the results to

the Belgian population.

2. Materials and methods

2.1. Population sample

2.1.1. Pre-school children

The first studied population included pre-school children living in Flanders (the northern,

Dutch-speaking region of Belgium that includes approximately 60% of the total Belgian

population). Representative samples of Flemish pre-school children aged 2.5 to 6.5 years

were selected on the basis of random cluster sampling at school level, stratified by province

and age. Details about the sampling strategy and the representativeness of the study sample

have been described by Huybrechts et al (2006; 2007). A total of 2095 children were enrolled

in this study. Their parents were asked to complete a food frequency questionnaire (FFQ), a

structured three-day estimated diet record (EDR), and a general questionnaire on socio-

demographic background, family composition, and child characteristics. Only the EDR data

were used to assess the PUFA intake, since the FFQ was developed to assess calcium intake

(Huybrechts et al, 2006).

Overall, 1052 subjects returned an EDR completed between October 2002 and February 2003.

Only EDRs containing sufficiently detailed descriptions of the food products and the portion

sizes consumed were used for analysis. The EDRs of 26 children were excluded due to

missing data. Of the 1026 remaining subjects, 696 completed three-day diaries, 208

completed two-day diaries, and 122 completed only a one–day diary (Huybrechts et al, 2006;

Huybrechts & De Henauw, 2007). In this study, only the data of children for whom three-day


44

diaries were available were used, since data from one day do not provide an accurate picture

of usual intake on the individual level, due to within-individual diet variation (Willett, 1998).

Since gender or age information was missing for 35 children, the EDRs of 661 children (338

boys and 323 girls) were included in the analyses.

2.1.2. Adolescents

The second subgroup of the population included Flemish adolescents. The survey was based

on a sample of 341 adolescents (129 boys and 212 girls), aged 13–18 years, from the region of

Ghent (± 250,000 inhabitants), a city in Flanders. Their food consumption data were collected

between March and May 1997 (Matthys et al, 2003). Different educational options of the

adolescents – ‘classical’ education (mainly theoretical courses) and vocational training (based

on practical skills) - were represented in the sample. A seven-day EDR (semi-structured

diary) was used to collect food consumption data. Instructions for the completion of the

diary and regular checks for quality and completeness of the diaries were carried out by

experienced dieticians. These data are the most recent available food consumption data on a

seven-day basis for Flemish adolescents.

2.1.3. Young women

The data source for the Flemish young women is a large epidemiological survey that

included women aged 18 to 39 years. A number of 4000 women were randomly selected

from the population register of Ghent, in Flanders. They were invited to participate in the

study by postal mail and were asked to reply by use of an enclosed postcard. Not less than

2634 subjects declined the invitation to participate and 424 invitation letters were declared

undeliverable by postal service. Women who were pregnant, who had moved to another

city, who did not speak Dutch, who were unable to come to the research centre, or who were

unable to volunteer within the planned period of the fieldwork, were excluded from the

study. In total, dietary data were collected for 641 women. The primary objective of the

survey was to evaluate food and nutrient intake (in particular iron) in a group of Flemish

women at reproductive age. Dietary assessment was done on the basis of a two-day EDR,

using diaries with a semi-structured, open-entry format. In addition, a general socio-

demographic questionnaire was administered. Data were gathered during the year 2002

(January 29th until December 22nd), thereby covering the different seasons. Only the data

from the two-day EDR are used in this study for assessing the intake of total fat and

important dietary fatty acids.


45

The three studies were approved by the Ethical Committee of the Ghent University Hospital.

Additional information about the pre-school children, adolescent, and young women

population has been previously published by Huybrechts et al (2006; 2007), Matthys et al

(2003; 2006), and Pynaert et al (2005), respectively. Experienced dieticians used a

standardized protocol, including a manual on household weights and measures to convert

the estimated amounts in the EDRs into weights (Belgian Health Council, 2005). None of the

individuals included in the studies reported the intake of supplements containing PUFAs.

2.2. Food composition database

The total fat content of the foods present in the food consumption databases of the three

studied population groups was obtained from the Belgian and Dutch food composition

database (FCDB) (NEVO Foundation, 2001; NUBEL, 1999). In the food consumption database

of the Flemish pre-school children, 726 of the 936 food items (77.6%) contained fat. In the

database of the adolescents and young women this was 70.6% (527 of the 745 food items) and

75.1% (700 of the 1063 food items), respectively.

To determine the PUFA concentration of the food items, a specific FCDB was developed that

included the concentrations of linoleic acid (LA), α-linolenic acid (LNA), arachidonic acid

(AA), eicosapentaenoic acid (EPA), docosapentaenoic acid (DPA), and docosahexaenoic acid

(DHA) as they were found in international literature sources. Eight publicly available FCDBs

were used, and, in order of the number of items that were used, they were:

1. the Dutch FCDB (NEVO Foundation, 2006);

2. an extended French FCDB (Astorg et al, 2004);

3. the USDA National Nutrient Database (US Department of Agriculture and Agricultural

Research Service, 2005);

4. the British McCance & Widdowson’s FCDB (Food Standards Agency, 2002);

5. the Finnish FCDB (National Public Health Institute of Finland, 2004);

6. the Danish FCDB (Danish Institute for Food and Veterinary Research, 2005);

7. the Canadian Nutrient File (Health Canada, 2005); and

8. the German FCDB (Souci et al, 2000).

Furthermore, food composition information was obtained from food producers for specific

food items, i.e. margarine, cheese, and dressings. For 46 composite food items, the fatty acid


46

(FA) composition was calculated using local recipes describing the different ingredients and

their proportions, as well as the FA composition of the ingredients as found in one of the

FCDBs. Detailed FA profiles of the fat-containing food items were then calculated by

calculating the proportional share of each FA in the total fat content as reported in the newly

compiled database and applying this proportional share to the total fat content of the food as

listed in the Belgian and Dutch FCDB (NEVO Foundation, 2001; NUBEL, 1999).

For vitamin D, data of the Dutch FCDB were used (NEVO Foundation, 2001), completed

with data from the Danish FCDB (Danish Institute for Food and Veterinary Research, 2005)

describing the vitamin D concentration in five seafood items for which data were lacking in

the Dutch data source. In total, 307 of the 745 food items (41.2%) in the adolescent food

consumption database contained vitamin D.

2.3. Statistics

All statistical analyses were done using SPSS software version 12.0 (SPSS, Inc., Chicago, IL,

USA). The average PUFA intakes were expressed in absolute amounts ((m)g/day) and in

percentage of the total energy intake (%E). Normality of the intake distributions was tested

using the Kolmogorov-Smirnov test (p < 0.01). A Mann-Whitney-U test was used to

determine differences in vitamin D (only for adolescents) and PUFA intakes between boys

and girls in the dataset of the pre-school children and the adolescents, as well as between

different age groups in the pre-school children dataset (p < 0.01).

Percentages of individual PUFAs provided by the different foods were calculated as

population proportions, as defined by Krebs-Smith et al (1989). The population proportion

was calculated by summing the amount of a FA from a certain food item for all individuals

and then dividing this number by the sum of that FA from all food items for all individuals.

The formula for determining the population proportion is:

∑

∑

=

== n

ii

n

ii

T

FproportionPopulation

1

1


47

Where Fi = the amount of the fatty acid contributed by the particular food group for the ith

individual,

Ti = the total amount of the fatty acid from all food for the ith individual,

n = the number of individuals in the total study population.

Using the adolescent food consumption data, the percentages were also calculated as mean

proportions. The mean proportion of a FA from a certain food item for all individuals is

determined by first calculating the contribution of a food item to the intake of that FA for

each person and then taking an arithmetic mean of all proportions (Krebs-Smith et al, 1989).

The formula for determining the mean proportion is:

n

TFproportionMean

n

iii∑

== 1)/(

The food items were grouped in 36 subgroups and 8 major groups, according to Astorg et al

(2004). A Mann-Whitney-U test was used to look for significant differences between boys

and girls in the contributions of different food groups to the nutrient intake (calculated as

mean proportion).

2.3.1. Pre-school children

The average PUFA intakes were calculated as the mean of the three-day period. The usual

PUFA intakes were computed using the NUSSER-method, a statistical method developed at

Iowa State University, based on the advice of the American Institute of Medicine (IOM) with

respect to the need to determine the distribution of usual nutrient intakes when assessing

diets of population groups in relation to the recommended levels (Institute of Medicine,

2005). The NUSSER-method estimates the usual intake distributions by accounting for

within-individual variation in nutrient intakes, while requiring relatively few days of intake

data per individual (Nusser et al, 1996). C-side software was used for the NUSSER-method

(Iowa State University, 2006). Moreover, exploratory calculations were done to assess the

amount of seafood that should be consumed by the pre-school children to achieve the

recommended amount of LC n-3 PUFAs.


48

2.3.2. Adolescents

The average vitamin D and PUFA intakes were calculated as the mean of the seven-day

period. For this subpopulation, no extra analyses were done to account for the within-

individual variation. In contrast, some additional analyses were done to study the

importance of seafood as a nutrient provider of vitamin D (in µg/d) and EPA plus DHA (in

mg/d). The study population was divided in tertiles based on the intake of these considered

nutrients. This division was done separately for boys and girls. A Kruskal-Wallis test was

used to test the importance of the different food groups between the tertiles (p < 0.01).

2.3.3. Young women

The average intakes of individual PUFAs of the group of young women were calculated

based on the mean of two days per individual, without correction for within-individual

variation.

2.4. Evaluation of the assessed intakes

To evaluate the intakes of the individual PUFAs and vitamin D (only assessed for

adolescents), the recommendations formulated by the Belgian Health Council were applied

(Belgian Health Council, 2007). The recommendations for PUFAs for children older than

three years of age on the one hand and for adults on the other hand are given in Table II.1,

expressed in %E. The PUFA recommendations for adults were applied to evaluate the

intakes of adolescents and young women. For vitamin D, the Belgian Health Council has

formulated a recommended daily intake range for adolescents and adults equal to 2.5 to 10

µg.


49

Table II.1: Recommendations for PUFAs for children older than three years of age and adults (expressed as percentage of the total energy intake (%E)) (Belgian Health Council, 2007) Recommendations (%E) Children older than three years of age Adults Total PUFAs > 8.0 5.3 - 10.0 LA 2 - 5 > 2.0 AA 0.10 - 0.25 - Total n-6 PUFAs - 4.0 – 8.0 LNA 0.45 - 1.50 > 1.0 EPA 0.05 - 0.15 - DHA 0.10 - 0.40 - EPA + DHA - > 0.3 Total n-3 PUFAs - 1.3 – 2.0

3. Results

3.1. Seafood consumption

Being of special interest in this PhD-thesis, the consumption of total seafood for the three

studied population groups is reported here. Most details are provided on the seafood

consumption data of the adolescents, since these data were used further on in the evaluation

of risk and benefits related to seafood consumption (chapter IV). Moreover, for the two other

population groups, consumption data were only available on two or three days, which is not

sufficient to give a good picture of episodically consumed food items such as seafood.

For the pre-school children it was found that only 207 of the 661 (31.3%) children did

consume seafood during the three days included in the study, with only 50 children

consuming fatty fish (salmon being the most important species). The mean seafood

consumption over the three days was 8.6 g/d for the total population sample and 27.4 g/d

for the subsample who consumed seafood.

The seafood consumption of the total group of studied adolescent population and the

seafood consumers only is summarized in Table II.2. From the 341 respondents, 63.9% did

consume seafood during the week of the study, respectively 81 boys and 137 girls (Table

II.2). The weekly amount of seafood consumption for boys was higher than for girls, but not

significantly different. The mean seafood eating occasions per week was 1.14, with a

maximum of five times a week. In total, 32 different seafood species and two seafood

products (caviar and surimi) were consumed by the adolescents. The most important species


50

were cod, saithe & Alaska Pollack, and salmon, accounting for more than half of the amount

of seafood consumed.

Table II.2: Seafood consumption of the studied adolescent population (g/week) and the subpopulation of seafood consumers only Seafood consumption (g/week) Whole population Seafood consumers only All

(n=341) Boys

(n=129) Girls

(n=212) All

(n=218; 64%) Boys

(n=81; 63%) Girls

(n=137; 65%) Mean 106.8 119.5 99.0 167.0 190.2 153.3 25th percentile 0.0 0.0 0.0 65.0 80.7 50.0 Median 55.0 70.0 48.8 148.0 150.0 125.0 75th percentile 183.0 184.5 184.0 221.0 226.3 219.5

Only 301 of the 641 (47.0%) of the Flemish women in the dataset consumed seafood during

the two-day study period. The average seafood consumption (calculated as a mean of the

two days) was 27.0 g/d for the whole group and 57.5 g/d for the seafood consumers only.

3.2. Assessed PUFA intakes

The intakes of LA, LNA, AA, EPA, DHA, total n-6 PUFA, and total n-3 PUFA for Belgian

pre-school children, adolescents, and young women and the corresponding

recommendations are shown in the figures below (Fig. II.1-3). The intakes as presented in the

histograms were not corrected for within-individual variation.

Fig. II.1 shows the intake of LA and LNA of the three studied population groups. The

resulting LA over LNA ratio is quite high: being 9.1 for the pre-school children and

adolescents and 8.7 for the young women.


51

Fig. II.1 Histograms of the LA (left) and LNA (right) intake of pre-school children (orange), adolescents (blue), and young women (green), as well as the recommended intake (red), expressed as percentage of the total energy

intake (%E)

Fig. II.2 shows the intake of AA only for the pre-school children as well as the intake of the

sum of LA and AA (further referred to as total n-6 PUFA intake or Σn-6PUFA) for the three

different subpopulations studied. The corresponding recommendation is also indicated in

the histograms, apart of the recommendation for total n-6 PUFA intake for pre-school

children because it is not specified by the Belgian Health Council (Table II.1). In general, the

contribution of AA to the ∑n-6PUFA was very low.

Fig. II.2 Histograms of the AA (for pre-school children only) and Σn-6PUFA (LA+AA) intake of pre-school children (orange), adolescents (blue), and young women (green), as well as the recommended intake (red),

expressed as percentage of the total energy intake (%E)


52

In Fig. II.3 the intake of EPA and DHA is represented separately for the pre-school children,

since a separate recommendation was formulated for both fatty acids. In contrast, for

adolescents and young women the intake is expressed as the sum of both fatty acids to

permit an easy comparison with the recommendations.

Fig. II.3 Histograms of the intake of EPA and DHA separately for pre-school children (orange) and the sum of

EPA and DHA for adolescents (blue) and young women (green), as well as the recommended intake (red), expressed as percentage of the total energy intake (%E)

Fig. II.4 shows the intake of the sum of LNA, EPA, DPA, and DHA (further referred to as

total n-3 PUFA intake or Σn-3PUFA) for the adolescents and the young women. This was not

shown for the pre-school children because no recommended intake was specified for them.

Fig. II.4 Histograms of the Σn-3PUFA (LNA+EPA+DPA+DHA) intake for adolescents (blue) and young

women (green), as well as the recommended intake (red), expressed as percentage of the total energy intake (%E)

Table II.3 shows the mean Σn-6PUFA/Σn-3PUFA ratio of the three studied subpopulations.

Table II.3: The mean Σn-6PUFA/Σn-3PUFA ratio for the three studied subpopulations Mean Σn-6PUFA/Σn-3PUFA ratio Pre-school children 8.5 Adolescents 8.1 Young women 7.8


53

In most cases, the PUFA intake distributions of the intakes were skewed to the right.

Moreover, a considerable part of the studied population groups had during the study period

an intake of LC PUFAs equal to zero. Normality was tested for the intake distributions

expressed in mg/day as well as in %E. In the group of pre-school children, only the intake

distribution of LA and ∑n-6PUFA expressed in %E, of the LA/LNA ratio, and of the ∑n-

6PUFA/∑n-3PUFA ratio were normally distributed. For the adolescents, only the intake

distributions of LA and ∑n-6PUFA expressed in mg and in %E, of LNA expressed in mg, and

of the ∑n-6PUFA/∑n-3PUFA ratio seemed to be normally distributed. In the group of young

women, only the intake distributions of LA and ∑n-6PUFA expressed in %E were normally

distributed.

3.2.1. The PUFA intakes of the pre-school children: additional results

The NUSSER-method applied to assess the usual PUFA intakes of the pre-school children

turned out to be only applicable for LA, LNA, Σn-6PUFA, and Σn-3PUFA, but not for

individual LC PUFAs: AA, EPA, DPA, and DHA (data not shown). The reason is that the

intake of these PUFAs over the three days was equal to zero for many children (Fig. II.2 and

II.3). For none of the calculated PUFA intakes significant differences were found between

boys and girls, neither between the children younger and older than four years of age.

Exploratory calculations were performed to assess the amount of seafood that should be

consumed to achieve the recommended amount of LC n-3 PUFAs. Starting from the Belgian

nutrient recommendations for EPA and DHA and a mean energy intake of 6543 kJ for boys

and 5757 kJ for girls (Huybrechts & De Henauw, 2007):

- the EPA intake has to be 608-1824 and 535-1605 mg/week, respectively for boys and

girls,

- the DHA intake has to be 1216-4865 and 1070-4281 mg/week, respectively for boys and

girls.

Based on the food composition data, it was calculated that two portions of 50 gram fatty fish

(e.g. mackerel, sardines, salmon) per week are needed to fulfil the requirements, assuming

that seafood is the only source of EPA and DHA. On the basis of the three-day EDRs, it was

found that only 50 of the 661 children consumed fatty fish (salmon being the most important

specie).


54

3.2.2. The vitamin D and PUFA intakes of the adolescents: additional results

The adolescents’ intakes of vitamin D (in µg/day as in µg/kJ/day) are given in Table II.4,

separately for boys and girls.

Table II.4: Intake of vitamin D (µg/day and µg/kJ/day) for Flemish adolescents Mean 5th percentile Median 95th percentile Vitamin D ♂ (µg/d) 4.0 1.8 3.6 7.6 Vitamin D ♂ (µg/kJ/d) 0.37 0.19 0.35 0.64 Vitamin D ♀ (µg/d) 2.5 1.3 2.5 4.9 Vitamin D ♀ (µg/kJ/d) 0.35 0.18 0.33 0.56

None of the vitamin D intake distributions were normally distributed. The vitamin D intake

was significantly higher for boys compared to girls when expressed in µg/day; after

dividing the vitamin D intake by the energy intake, no significant difference was found

between both genders.

For the PUFA intakes, significant differences were found between boys and girls for LA,

LNA, AA, ∑n-6PUFA, and ∑n-3PUFA when expressed in g/d, with the intakes of the boys

higher than of the girls (data not shown). When comparing the intakes expressed as %E no

significant differences were detected. This is explainable by the correction made by using the

energy intake that was significantly higher in boys (10625 kJ) as compared with girls (8030

kJ) (Matthys et al, 2003).

3.3. Food sources of the individuals PUFAs

3.3.1. The PUFA food sources for the pre-school children

Fig. II.5 shows the dietary sources of the different PUFAs for the Flemish pre-school children

calculated as population proportions.


55

0%

20%

40%

60%

80%

100%

LA LNA AA EPA DPA DHA

Sweet products

Miscellaneous

Meat, poultry & eggs

Fruits & vegetables

Fish & seafood

Fats & oils

Dairy products

Cereal products

Fig. II.5 Food sources of the different PUFAs for the studied pre-school children (% of the total intake of each FA brought by each food group for the whole population. The group ‘Miscellaneous’ contains salty & other snacks,

soy drinks, vegetarian substitute, and spices & condiments)

- It is remarkable that the contribution of sweet products (including biscuits, chocolate

products, pastry & desserts, and sugar & sweets) to the LA intake was as high as the

contribution of fats & oils (including fatty sauces, margarines, vegetable oils, and mixed

fats). Within the group of sweet products, the most important contributor was chocolate

products (11.6%), followed by biscuits (8.3%). Two other important LA contributors were

cereal products (with bread as most important contributor) and meat, poultry & eggs.

- The contribution pattern of the different food groups for LNA was quite similar to the

one found for LA, with almost one third from the consumption of fats & oils, with

margarines being the most important subgroup (Fig. II.5).

- The AA intake was mainly contributed by meat, poultry & eggs (32.6% by poultry and

34.3% by meat & meat dishes). The second important contributor was fish & seafood

(11.1% from lean fish and 7.4% by molluscs & crustaceans).

- For the LC n-3 PUFAs, fish & seafood was the major source, with fatty fish being the

most important subgroup (a contribution of 53.4%, 42.8%, and 48.1% for EPA, DPA, and

DHA, respectively). A substantial part of the DPA intake came from poultry (23.5%) and

meat & meat dishes (18.1%). Owing to fortification of margarines and the use of eggs in

the preparation of certain sweet products and snacks, these products also contributed to

the LC n-3 PUFA intake (a contribution of 1.2%, 0.1%, and 1.3% by sweet products and of

1.5%, 0.5%, and 1.9% by snacks for EPA, DPA, and DHA, respectively).


56

3.3.2. The vitamin D and PUFA food sources for the adolescents

Table II.5 shows the proportional contribution of the different food groups to vitamin D and PUFA intake of the studied adolescents, calculated as population proportions. Table II.5: Contribution of food groups to vitamin D (VitD), n-6 and n-3 PUFA intakes (% of the total intake of each FA brought by each food group for the whole population) for the Flemish adolescents

Food groups VitD LA LNA AA EPA DPA DHA ∑n-6PUFA ∑n-3PUFA Bread & rusks 0.12 18.21 12.97 0.00 0.00 0.00 0.00 18.08 11.45 Breakfast Cereals 0.00 0.51 0.21 0.00 0.00 0.00 0.00 0.51 0.18 Cereal based dishes 1.30 1.25 1.77 0.75 0.00 0.00 0.17 1.24 1.57 Pasta, rice & other cereals 0.73 1.77 1.12 0.32 0.02 0.00 0.00 1.76 0.99 Total cereal products 2.16 21.74 16.06 1.07 0.02 0.00 0.17 21.59 14.18 Butter 0.69 0.16 0.64 0.68 0.00 0.00 0.00 0.16 0.56 Cheese 4.90 1.02 3.56 2.22 0.00 0.00 0.00 1.03 3.14 Cream 0.34 0.05 0.17 0.00 0.00 0.00 0.00 0.05 0.15 Milk 0.82 0.37 1.08 0.42 0.00 0.00 0.00 0.37 0.95 Yogurts 0.24 0.05 0.16 0.02 0.00 0.00 0.00 0.05 0.14 Total dairy products 7.00 1.65 5.60 3.34 0.00 0.00 0.00 1.66 4.94 Fatty sauces 0.44 10.54 32.86 0.00 0.44 0.00 0.00 10.47 29.01 Margarines 36.43 19.23 15.25 0.00 0.00 0.00 0.00 19.09 13.45 Mixed fats 0.02 5.15 0.08 0.02 0.00 0.00 0.00 5.11 0.07 Vegetable oils 0.00 8.55 0.82 0.00 0.00 0.00 0.00 8.49 0.73 Total fats & oils 36.89 43.47 49.00 0.02 0.44 0.00 0.00 43.16 43.25 Fatty fish 8.28 0.21 0.44 1.23 42.36 40.68 29.87 0.21 4.47 Fish products 1.45 0.21 0.07 0.47 4.62 2.28 4.02 0.21 0.54 Half-fatty fish 0.58 0.03 0.05 0.73 4.28 3.51 9.39 0.04 0.90 Lean fish 2.57 0.01 0.03 2.09 17.17 8.51 16.56 0.02 1.90 Molluscs & crustaceans 0.27 0.01 0.03 3.16 15.68 4.29 5.55 0.03 1.02 Total fish & seafood 13.15 0.46 0.62 7.68 84.11 59.27 65.40 0.51 8.83 Fruits 0.00 0.03 0.31 0.00 0.00 0.00 0.00 0.03 0.28 Legumes 0.00 0.02 0.13 0.00 0.00 0.00 0.00 0.02 0.12 Nuts & seeds 0.00 1.42 0.45 0.00 0.02 0.00 0.00 1.41 0.40 Potatoes 12.72 0.06 0.21 0.00 0.00 0.00 0.00 0.06 0.18 Soups 0.39 0.10 0.06 0.03 0.00 0.00 0.00 0.10 0.05 Vegetables 0.00 0.19 0.68 0.02 0.06 0.00 0.00 0.19 0.60 Total fruits & vegetables 13.11 1.83 1.84 0.04 0.08 0.00 0.00 1.81 1.63 Eggs 5.25 1.29 0.41 20.43 0.34 10.22 8.54 1.42 1.10 Meat & meat dishes 9.65 7.61 8.25 35.59 4.29 6.78 9.03 7.80 8.15 Poultry 3.00 2.26 1.72 25.11 5.36 19.60 10.35 2.42 2.66 Total meat, poultry & eggs 17.90 11.16 10.38 81.13 9.99 36.60 27.92 11.65 11.91 Biscuits 2.38 3.51 2.68 0.27 0.00 0.46 0.07 3.49 2.38 Chocolate products 0.53 8.39 5.77 0.00 0.00 0.00 0.00 8.34 5.09 Pastry & desserts 5.89 2.73 3.00 4.32 0.06 0.00 1.07 2.74 2.72 Sugar & sweets 0.00 0.03 0.02 0.00 0.00 0.00 0.00 0.03 0.02 Total sweet products 8.80 14.66 11.47 4.60 0.06 0.46 1.14 14.59 10.21 Miscellaneous 0.00 0.01 0.02 0.00 0.00 0.00 0.00 0.01 0.02 Salty snacks 0.02 2.99 1.10 0.00 0.00 0.57 0.08 2.97 0.99 Snacks 0.97 1.64 3.41 2.13 5.19 3.10 5.29 1.65 3.60 Spices & condiments 0.00 0.09 0.32 0.00 0.00 0.00 0.00 0.09 0.28 Vegetarian substitute 0.00 0.31 0.18 0.00 0.10 0.00 0.00 0.31 0.16 Total miscellaneous 0.99 5.04 5.03 2.13 5.30 3.67 5.36 5.02 5.05 ∑n-6PUFA= LA+AA; ∑n-3PUFA= LNA+EPA+DPA+DHA


57

- The major sources of vitamin D were fats & oils, with margarines as most important

subgroup. Other important contributors were meat, poultry & eggs, fish & seafood, and

prepared potatoes (e.g. mashed potatoes, containing eggs and margarine).

- Fats & oils were also the main source for LA, followed by cereal products. The

contribution of sweet products to the LA intake was less important than for the pre-

school children (Fig. II.5). In the group of fats & oils, margarines, and fatty sauces

(dressings, etc.) were the major sources (Table II.5).

- For LNA, again a quite similar result as for LA was found, with fats & oils being the

major contributors, followed by cereal products. In contrast, where dairy products were

of negligible importance for the LA intake, they counted for 5.6% for the LNA-intake,

with cheese being the most important subgroup. In the group of cereal products, bread &

rusks were most important.

- The intake of AA was mainly contributed by meat, poultry & eggs.

- For the three different LC n-3 PUFAs, fish & seafood contributed for 84.1%, 59.3%, and

64.4% respectively for EPA, DPA, and DHA. The most important subgroup was fatty

fish. For EPA, molluscs and crustacean were also quite important. A substantial part of

the DPA intake was contributed by poultry, meat & meat dishes, and eggs. They

contributed also to the DHA intake, with poultry as major subgroup. Consequently, we

can conclude that for the LC PUFAs the food sources are similar for the pre-school

children as for the adolescents.

The contribution of each food group to the nutrient intake was also calculated on individual

level (mean proportion) and then compared between boys and girls (data not shown). No

relevant differences were found on that level. More detailed information about the actual

consumption of the different food groups is previously published (Matthys et al, 2006).

Table II.6 indicates the energy and fat intake and the consumption of some relevant food

items for the tertiles based on the vitamin D and EPA&DHA intake. The consumption of the

food items is expressed in g/d as in contribution to the total energy intake (%E); the latter to

correct for the overall energy intake of the individuals.


58

Table II.6: Consumption of different food items (in g/day and in contribution to the total energy intake (%E)) for the different tertiles based on the vitamin D intake and the EPA&DHA intake, separately for boys and girls * ♂ (n=120) ♀ (n=212) Tertiles of vitamin D intake (µg/d) <3.00 [3.00-4.33] >4.33 < 2.25 [2.25-3.05] >3.05 Energy intake kJ 9761 10553 11937 6847 8083 9122

g/d 89.6 103.5 122.8 61.2 77.3 92.0 Fat intake %E 34.5 36.9 38.7 33.4 36.1 38.0 g/d 8.3 16.7 36.2 6.4 11.9 20 Margarine %E 2.4 4.2 7.4 2.5 4 5.3 g/d 7.8 12.7 16.6 6.4 8.5 11.6 Total fats & oils %E 2.3 3.4 3.7 2.7 3 3.4 g/d 21.2 23.3 27.7 18.8 19.4 29.7 Lean fish %E 0.9 0.8 0.8 1 1 1.3 g/d 16.5 18.6 13.7 8.2 10.4 10.6 Half-fatty fish %E 1 1.1 0.5 0.6 0.7 0.5 g/d 7.6 8.2 22.8 5.5 7.2 22.4 Fatty fish %E 0.6 0.7 1.7 0.9 0.9 2.8 g/d 14.1 15.7 20.4 11.4 13 18.8 Total fish & seafood %E 0.7 0.9 1 0.8 0.8 1.3

Tertiles of EPA&DHA intake (mg/d) <70 [70-180] >180 <54 [54-143.6] >143.6 Energy intake kJ 10848 10537 10864 7740 7924 8377

g/d 105.9 104.3 105.2 73.4 75.9 80.9 Fat intake %E 36.7 37.3 36.2 35.4 35.8 36.3 g/d 6.7 17.8 31 8.3 14.4 26.2 Lean fish %E 0.2 0.6 1 - 0.7 1.2 g/d 2.9 14.5 17.5 3.9 8.5 11.7 Half-fatty fish %E 0.1 0.8 0.9 0.2 0.6 0.7 g/d 8.2 3.6 17.8 2.1 4.9 18.7 Fatty fish %E 0.5 0.3 1.4 0.3 0.7 2.3 g/d 3.6 6.5 12.6 3.8 7 15.9 Molluscs & crustacean %E 0.1 0.2 0.4 0.2 0.4 0.7 g/d 15 18.5 18.5 9.7 14.9 14.4 Fish products %E 1.2 1.4 1.3 0.9 1.3 1.3 g/d 8.3 12.9 20.1 6 10 18.6 Total fish & seafood %E 0.5 0.6 1.1 0.5 0.7 1.3

For the figures indicated in bold, a significant difference between the tertiles was found (p < 0.01) * The amounts consumed of the different food items reported in g/day and in %E are calculated as the mean only for those adolescents that consumed the food item during the week of the study and not as a mean for all members of the tertile. In other words, it is a mean of the consumer population only. As a result, the sum of the mean consumption of the different fish types is not equal to the consumption of total fish & seafood, since a different number of consumers is accounted for.

- For both genders, higher energy and fat intake as well as higher consumption of

margarine and total fats & oils was found for the higher vitamin D tertiles. This was also

the case for fatty fish, expressed in g/day (Table II.6). Moreover, for the girls, significant

differences were found for the consumption of pastry & desserts, potatoes, and poultry

(in g/d), with a higher consumption for the higher vitamin D tertiles (data not shown).

- When considering the EPA&DHA tertiles, significant differences were only found for

food items belonging to the fish & seafood group, showing that higher EPA&DHA intake

was related with higher fish & seafood intake; the intake of energy and fat did not differ


59

significantly. For non-mentioned food groups (e.g. fruits & vegetables), no significant

differences were found over the different tertiles.

3.2.3. The PUFA food sources for the young women

Table II.7 shows the proportional contribution of the different food groups to PUFA intake

of the studied young women, calculated as population proportions.

As for the adolescents, fats & oils were the main sources for LA and especially for LNA,

followed by cereal products. In the group of fats & oils, margarines and fatty sauces

(dressings, etc.) were the major subgroups. Compared to the adolescents, a more important

contribution of nuts and seeds to the LA and LNA intake was found for the young women.

Again, the intake of AA was mainly contributed in the diet by meat, poultry & eggs,

however, for this population group eggs was the most important subgroup. As found for the

two other studied population groups, fish & seafood are the most important food sources of

EPA, DPA, and DHA, even more important for the young women than for the adolescents.


60

Table II.7: Contribution of food groups to n-6 and n-3 PUFA intakes (% of the total intake of each FA brought by each food group for the whole population sample) for the Flemish young women Food Groups LA LNA AA EPA DPA DHA ∑n-6PUFA ∑n-3PUFA Bread & rusks 12.9 9.9 0.0 0.0 0.0 0.0 12.8 8.5 Breakfast Cereals 1.1 0.5 0.2 0.0 0.0 0.0 1.1 0.4 Cereal based dishes 0.8 1.4 0.7 0.0 0.0 0.1 0.8 1.2 Pasta, rice & other cereals 1.6 1.3 0.6 0.4 0.0 0.2 1.6 1.2 Total cereal products 16.4 13.1 1.6 0.4 0.0 0.3 16.3 11.3 Butter 0.3 0.9 1.2 0.0 0.0 0.0 0.3 0.8 Cheese 1.2 4.4 4.3 0.0 0.0 0.0 1.3 3.7 Cream 0.1 0.5 0.0 0.0 0.0 0.0 0.1 0.4 Milk 0.2 0.6 0.4 0.0 0.0 0.0 0.2 0.6 Yogurts 0.1 0.2 0.0 0.0 0.0 0.0 0.1 0.2 Total dairy products 1.9 6.6 5.9 0.0 0.0 0.0 2.0 5.6 Fatty sauces 11.3 25.1 0.0 0.1 0.1 0.0 11.2 21.6 Margarines 15.7 17.7 0.0 0.9 0.0 0.5 15.6 15.3 Mixed fats 0.1 0.2 0.0 0.0 0.0 0.0 0.1 0.1 Vegetable oils 4.3 1.8 0.0 0.0 0.0 0.0 4.3 1.6 Total fats & oils 31.4 44.7 0.0 1.1 0.1 0.5 31.2 38.6 Fatty fish 0.4 0.6 5.0 47.4 49.8 48.1 0.4 7.3 Fish products 0.4 0.1 0.4 3.2 1.9 3.0 0.4 0.5 Half-fatty fish 0.0 0.0 1.2 5.2 3.8 6.6 0.0 0.8 Lean fish 0.0 0.0 2.2 7.6 4.5 9.5 0.0 1.2 Molluscs & crustaceans 0.0 0.1 8.9 23.9 6.1 12.9 0.1 2.3 Total fish & seafood 0.8 0.9 17.8 87.3 66.0 80.0 0.9 12.1 Fruits 0.0 0.4 0.0 0.0 0.0 0.0 0.0 0.4 Legumes 0.1 0.3 0.0 0.0 0.0 0.0 0.1 0.3 Nuts & seeds 6.6 3.5 0.0 0.6 0.0 0.0 6.6 3.1 Potatoes 5.5 1.7 0.0 0.0 0.0 0.0 5.5 1.5 Soups 0.6 0.6 0.8 0.1 0.0 0.2 0.6 0.5 Vegetables 0.6 1.8 0.0 0.1 0.0 0.0 0.6 1.5 Total fruits & vegetables 13.4 8.4 0.8 0.8 0.0 0.2 13.4 7.2 Eggs 1.1 0.3 25.5 0.2 6.2 6.1 1.2 0.9 Meat & meat dishes 8.4 8.3 17.9 2.4 12.1 2.0 8.5 7.6 Poultry 1.6 1.2 15.3 2.7 12.9 3.7 1.6 1.7 Total meat, poultry & eggs 11.1 9.9 58.7 5.2 31.2 11.8 11.3 10.1 Biscuits 2.9 2.7 0.2 0.0 0.0 0.0 2.8 2.4 Chocolate products 5.4 3.7 0.0 0.0 0.0 0.0 5.4 3.2 Pastry & desserts 4.5 4.4 8.7 0.1 0.1 0.9 4.5 3.8 Sugar & sweets 0.1 0.0 0.0 0.0 0.0 0.0 0.1 0.0 Total sweet products 12.8 10.8 9.0 0.1 0.1 0.9 12.8 9.3 Dietetic products 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Miscellaneous 0.1 0.0 0.0 0.0 0.0 0.0 0.1 0.0 Salty snacks 2.2 0.7 0.0 0.0 0.2 0.0 2.1 0.6 Snacks 6.8 2.9 6.2 4.3 2.3 6.2 6.8 3.2 Soydrink 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Spices & condiments 0.5 0.8 0.0 0.0 0.0 0.0 0.5 0.7 Vegetarian substitute 2.6 1.3 0.0 0.8 0.0 0.0 2.6 1.2 Total miscellaneous 12.1 5.8 6.2 5.1 2.5 6.2 12.1 5.7 ∑n-6PUFA= LA+AA; ∑n-3PUFA= LNA+EPA+DPA+DHA


61

4. Discussion

4.1. Methodological limitations

4.1.1. Food consumption data

A first methodological limitation was related to the food consumption data used. A

weakness of the EDR method is underreporting or underestimation of the food intake. In

general, food items rich in fat and/or carbohydrates (such as butter, sweet products, and

snacks) are reported less frequently and/or in smaller quantities than actually consumed.

This has an influence on the assessed nutrient intake. Moreover, the number of recorded

days is restricted for the food consumption dataset of the pre-school children and the young

women. A three-day and two-day EDR for the pre-school children and the young women,

respectively, does not necessarily reflect usual intake of individuals and is likely to lead to

more extreme high and low intakes, especially for nutrients abundantly present in foods that

are not consumed on a daily basis, such as seafood (Willett, 1998). This is illustrated in Fig.

II.1-4, indicating the smaller ranges for the assessed intakes of the adolescent population

compared to those of the studied pre-school children and young women. The main reason is

a high within- and between-individual variation for these nutrient intakes over different

days.

Nelson et al (1989) calculated the days needed to assess the individual PUFA intake with a

good correlation (r ≥ 0.9) between the observed and the true mean intake. At least 15, 18, 28,

and 30 days should be included for toddlers (1-4 years), male children (5-17 years), female

children (5-17 years), and women older than 18 years, respectively. However, this is not

feasible in practice. As an alternative, different methodologies are developed to take into

account the within-individual variation. These methodologies include statistical modelling to

estimate long-term usual intake from food consumption data recorded on a limited number

of days (Dodd et al, 2006). Applying such statistical methodologies should be a more efficient

solution than asking respondents to report their consumption pattern during more than

seven days. It is known that increasing the number of days per study person leads to high

respondent burden and risks to decrease the quality of the reported information.


62

One of the existing modelling methods is the NUSSER-method, which was applied to the

dataset of pre-school children in order to assess usual PUFA intakes. However, the NUSSER-

method turned out to be inapplicable for these PUFAs for which a substantial part of the

population had a zero intake, being the major disadvantage of this method. Nevertheless, the

mean PUFA intakes on population level, resulting from the three-day EDRs did not differ

much from the usual nutrient PUFA estimates assessed by the NUSSER-method (data not

shown). This might be due to rather small within-individual variation of the consumption

pattern of young children. Recently, improved methods for estimating usual intake

distributions for episodically consumed foods are being developed, but were not yet

available on the moment of the study (Dodd et al, 2006). Nevertheless, it was assumed that

the three- and two-day EDRs yielded unbiased estimates on population level of the mean

food consumption and nutrient intakes. Of course, the data can not be used to evaluate the

intake on individual level.

In contrast to the datasets of the pre-school children and the young women, the adolescent

food consumption data describe the consumption during seven consecutive days. The

important advantage of these consumption data gathered over a longer period involves that

they allow assessing the intake of nutrients present in relatively few foods that are not eaten

on daily basis, such as seafood (Lamberg-Allardt, 2006). However, the disadvantage of this

dataset is that it concerns rather old data (1997).

4.1.2. Food composition data

Another limitation of this study is that no own analytical concentration data for vitamin D

and PUFA were at our disposition. Hence, several previously published data were used, but

the derivation of this data is unknown. Additionally, it is noticeable that in individual foods

PUFAs, in particular the LC PUFAs, are present in only small amounts, but accumulate to

significant levels of biological importance in the context of a whole diet. For most foods,

these low values often round down to zero when reported on a single decimal place and so it

is likely that PUFA concentrations will be consistently underreported in FCDBs (Mann et al,

2003) leading to an underestimation of the assessed PUFA intake. Furthermore, for some

items it was not evident to find the PUFA composition because new (highly processed or

imported) food items are available on the market on a regular basis. Quinoa (a South-


63

American crop primarily grown for its edible seeds which are prepared and eaten as a

cereal), for example, was only mentioned in the German FCDB (Souci et al, 2000). Eight

different FCDBs were necessary to determine the PUFA composition of all consumed food

items, which is not an ideal situation since different protocols can be hidden behind the data

and since the origin of the published data was hard to trace for some FCDBs.

Moreover, a very recent Australian study noted that the meat & poultry was a significant

source of LC n-3 PUFA (Howe et al, 2006). The authors stated that the modest amount of

PUFA in muscle tissue phospholipids of lean meat needs to be taken into account when

determining dietary PUFA intakes. The majority of LC n-3 PUFA in meat is DPA. According

to Howe et al (2006) the LC n-3 PUFA content of meat products are underestimated up to

now. Since the limitations of available data about the fatty acid content of these food items,

LC n-3 intake from meat, poultry & eggs, especially DPA intake, could have been

underestimated. Nevertheless, the Belgian recommendation considers only EPA and DHA

for the evaluation of LC n-3 PUFA intake (Belgian Health Council, 2007). Therefore, the

evaluation of the intake data against the Belgian recommendations is not influenced by the

possible underestimation of DPA during the intake assessment.

4.2. Evaluation of the intakes

4.2.1. Observations versus recommendations

4.2.1.1. LA and LNA

Fig. II.1 shows that the intake of LA of almost all individuals of the three studied population

groups exceeded the recommendation, indicating that the risk of LA deficiency is extremely

low. This confirmed the abundance of LA sources in the modern Western diet (Sontrop &

Campbell, 2006). Therefore, it can be concluded that no dietary shifts are needed to increase

the current LA intake. In contrast, a non-negligible part of the three considered population

groups had a LNA intake lower than the recommended level (Fig. II.1). Since the food

sources of LA and LNA are quite similar (Fig. II.5, Table II.5 and II.6), a general shift in the

consumption of a certain food group to increase the LNA intake will simultaneously have an

effect on the LA intake. Therefore, a shift between foods within a certain food group is

needed. The consumption of LNA rich foods by the Flemish population should be


64

encouraged, specifically by encouraging the consumption of food items with a lower

LA/LNA ratio e.g. rapeseed oil, linseed oil, soy bean oil, and walnut (oil), either by direct

consumption of these foods or by increased use of these ingredients in processed foods, such

as margarines and biscuits, which must be possible by new and highly-developed techniques

in food manufacturing. Considering margarine, food industries must try to keep the trans

fatty acid concentration of their products as low as possible, since these fatty acids are

known to increase the risk of coronary heart disease (Korver & Katan, 2006; Upritchard et al,

2005).

4.2.1.2. AA

Fig. II.2 shows that the AA intake of pre-school children is extremely low compared to the

recommendation. An in-depth discussion on the AA intake is given below (paragraph

4.2.2.1.). On the other hand, Fig. II.2 indicates that the mean ∑n-6 PUFA intake of the studied

adolescents and young women fall within the recommended range, i.e. 4-6 %E.

4.2.1.3. LC n-3 PUFA

Fig. II.3 and II.4 indicate clearly that the intakes LC n-3 PUFA (EPA and DHA) as well as the

∑n-3 PUFA fall well below the recommended intake. As such, it can be concluded that all

studied population groups have an important deficit for these fatty acids.

4.2.1.4. Flemish versus Belgian data

As stated in the introduction of this chapter, the available food consumption data used for

the PUFA intake assessment were limited to the Flemish population. Nevertheless, the

Belgian Health Interview Survey (HIS) (Scientific Institute of Public Health, 2006) and the

Belgian Food Consumption Survey (FCS) (De Vriese et al, 2006) provided data about seafood

consumption and these surveys had sampled through all inhabitants of Belgium. However,

both datasets are limited to persons over 15 years.

- Data of the HIS of 2004 indicated that the percentage of people eating at least as much

fish as recommended was higher in the Flemish region (66.0%) than in the Walloon

region (54.9%), and that the Brussels region had the highest percentage (67.0%).

Unfortunately, it was not indicated what was meant by ‘at least as much fish as

recommended’ and it was not clear whether only fish or also other seafood was

considered.


65

- In contrast, the data of the FCS indicated that almost 70% of the Belgian population

consumed less seafood than recommended. Here, a recommendation of 30 g/day was

applied. Despite of the different results, the FCS data also indicated that the seafood

consumption in the Flemish region is higher than the seafood consumption in the

Brussels and Walloon region.

Therefore, it can be assumed that the conclusion that the intake of LC n-3 PUFA is very low

compared to the recommendations is true for the Flemish region as well as for the other

Belgian regions.

4.2.1.5. Seafood consumption recommendations

Increased seafood consumption is one of the possibilities to increase the LC n-3 PUFA intake.

Currently, the Belgian Health Council advices the population to consume seafood one to two

times a week (Belgian Health Council, 2004b). Moreover, seafood has a prominent place in

the Flemish food triangle and the Walloon food pyramid. The Belgian Health Council

underlines the recommendation to consume different species from various origins.

Moreover, they advise pregnant and lactating women to consume only one portion of tuna

per week (Belgian Health Council, 2004a). In chapter IV of this PhD-thesis, it is investigated

whether this recommendation is sufficient to achieve the DRI for EPA and DHA and

whether an upper limit is needed related to contaminant intake.

Nevertheless, 36.1% of the studied adolescents did not consume any seafood in the week of

the study. An evaluation of the seafood consumption for the two other studied populations

is difficult since only short term consumption data were available. Nevertheless, a more

regular replacement of meat & meat products by seafood should, above an increased intake

of LC n-3 PUFAs and vitamin D, also be beneficial to decrease the too high saturated fatty

acid (SFA) intake of the studied populations, as previously reported (Huybrechts & De

Henauw, 2007; Matthys et al, 2006; Sioen et al, 2006a). Additionally, replacement of high-fat

cheese regularly consumed during bread meals, with seafood like mackerel and sardines can

lower the SFA intake and at the same time increase the intake of vitamin D and LC n-3

PUFAs. These arguments are also a counter-argument to the intake of supplements. While

LC n-3 PUFA supplements can seem to be an easy solution for low LC n-3 PUFA intake, the

presence of other nutrients in food sources of EPA and DHA must also be considered, such

as important amino-acids and trace elements together with a low concentration of SFA and

cholesterol.


66

4.2.1.6. Seafood consumption barriers

Nevertheless, one could identify different barriers to recommend increased seafood

consumption:

1. There are concerns that higher fatty fish consumption could impact on intakes of

contaminants such as dioxin-like compounds and heavy metals. But, currently, more and

more studies indicate that the potential benefits from LC n-3 PUFAs outweigh the risk of

exposure to toxicological levels of contaminants (American Heart Association et al, 2006;

Ruxton et al, 2004; SACN/COT, 2004). A more detailed evaluation of the risks and

benefits related to seafood consumption can be found in the following chapters of this

PhD-thesis.

2. The food choice of young people is mostly driven by taste, smell, and convenience, and

seafood does not have a high preference rate (Diehl, 1999). Additionally, in the case of

pre-school children and adolescents, it will be in the first place the decision of the parents

or school catering services whether or not seafood will be regularly placed on the menu.

Furthermore, although parent’s food consumption decisions are shaped by personal

norms or moral obligation to serve their children a nutritious meal, children’s preferences

and the desired compliance with these play a role as well (Kelly et al, 2006). Recent

empirical evidence suggests that the presence of young children in the family acts as an

important barrier to increased seafood consumption, particularly in countries with a

relatively weak seafood consumption tradition such as Belgium. The presence of young

children in Belgian households has been associated with a lower impact of external social

norms, i.e. parents’ interest and willingness to take into account the advice from external

information sources in their social environment such as medical sources, in determining

seafood consumption intention (Verbeke & Vackier, 2005).

3. Belgian adults’ awareness of both the fact that seafood is an important dietary source of

LC n-3 PUFA, and the fact that these PUFAs have a potential beneficial impact on human

health, is rather limited (Verbeke et al, 2005).

4. Last but not least, fish and/or seafood allergies or total dislike of seafood will influence

the actual consumption. In the latter case, other strategies, e.g. fortified foods and

supplements, can be used as alternatives to increase the LC n-3 PUFA intake.


67

4.2.2. Evaluation of the results and comparison with international data

In the paragraphs below, the intakes of the studied subgroups of the Flemish population are

compared to international data. Without doubt, different food composition data and

different methodologies to collect food consumption data were used in the different studies

hampering the comparability of the data. Moreover, non-negligible differences exist in the

age and the total energy intake of the population. Yet, it was worth to compare the results to

evaluate the PUFA intake data of the Flemish population groups in an international context.

4.2.2.1. PUFA intakes of pre-school children

The PUFA intakes of the pre-school children were compared to results from other studies:

1. Canadian children (n=84; 1.5-5 years), based on an FFQ;

2. Australian children (2-3 years (n=383) and 4-7 years (n=799)), based on one 24-h recall;

3. children in rural China (n=196; 1-5 years) based on three 24-h recalls (Table II.8).

Table II.8: PUFA intakes of pre-school children from different studies *

LA g/d

LA %E

LNA g/d

LNA %E

AA mg/d

AA %E

EPA mg/d

DPA mg/d

DHA mg/d

DHA %E

LA/ LNA

Flemish† 2.5-3y 6.71 4.29 0.81 0.51 17 0.01 22 10 43 0.03 9.1 Flemish† 4-6.5y 7.12 4.27 0.86 0.51 18 0.01 26 10 49 0.03 9.1 Australian ‡ 2-3y 6.10 - 0.68 - 16 - 10 5 24 - - Australian ‡ 4-7y 7.50 - 0.81 - 22 - 19 10 47 - - Canadian § 2y 9.04 - 2.02 - 260 - 57 - 95 - 5.2 Canadian § 3-5y 9.39 - 1.72 - 226 - 60 - 96 - 6.6 Chinese || 1-3y 2.08 2.9 0.28 0.4 55 0.08 - - 34 0.05 - Chinese || 4-5y 2.27 2.5 0.34 0.4 50 0.06 - - 23 0.02 - † This study; ‡ (Meyer et al, 2003); § (Innis et al, 2004); || (Barbarich et al, 2006) * Since values expressed as %E were not found in other studies for EPA and DPA, these were not given in the table.

- Comparison of the LA intakes illustrated that the intake of Flemish pre-school children

was similar to the intake of Australian, lower than that of Canadian, but much higher

than that of Chinese children (Table II.8).

- The LNA intake of Flemish pre-school children was - as for the LA intake - comparable to

that of Australian, but much lower than that of Canadian children. The LNA intake of

Chinese children seemed to be very low when expressed in g/d, but only a bit lower than


68

the Flemish intakes when expressed in %E, due to the low energy intake of rural Chinese

children.

- The rather high LA intake and the too low LNA intake of Flemish pre-school children

resulted in a high LA/LNA ratio, being much higher than the LA/LNA ratio of Canadian

children. This confirms that the consumption of food items with lower LA/LNA ratio by

Flemish pre-school children (and by the Flemish population in general) should be

stimulated and that food industries must be encouraged to lower the LA/LNA ratio of

their products, if possible.

- All Flemish pre-school children had an AA intake lower than the recommended lower

level. Different possible explanations or influencing factors can be put forward:

1. the overall diet truly contain very little AA,

2. the recommended AA intake range set by the Belgian Health Council is too high,

3. the AA concentrations in the foods were underestimated in the FCDB used.

Considering the second explanation, no specific AA recommendations are given by the

EU (Commission of the European Communities, 1993) or the American Institute of

Medicine (IOM) (Institute of Medicine, 2005), hence hampering a comparison. However,

the IOM indicates that 5 to 10 %E should come from n-6 PUFAs, with approximately 10%

from LC n-6 PUFAs. Considering the third explanation, literature appears to be

equivocal. An Australian study stated that AA values in FCDB are too high, leading to an

overestimation of the AA intake (Mann et al, 1995). In contrast, an American study

concluded that AA values in the American FCDB (version HB-8) were significantly lower

compared to their analytical results (Taber et al, 1998). The AA intakes of Flemish and

Australian pre-school children were comparable. That of Chinese children were more

than two times higher, and that of Canadian more than 10 times higher (chicken was the

major source of dietary AA for many of the Canadian children). Possible ways to increase

the current AA intake are increased consumption of poultry meat. Nevertheless, when

recommending increased meat consumption, attention must be paid not to increase the

overall saturated fatty acid intake.

- Canadian children had the largest LC n-3 PUFA intake being twice as high as the intake

of the Flemish ones (Table II.8). Exploratory calculations showed that stimulation of fatty

fish consumption among Flemish pre-school children is a possible solution to bridge the

gap between the intake and the recommendation.


69

4.2.2.2. Vitamin D and PUFA intakes of adolescents

The mean and median vitamin D intakes of the adolescent boys were situated within the

recommended range. In contrast, the results showed that half of the studied girls had an

intake lower than 2.5 µg/d. The most important dietary source of vitamin D for the

adolescents was margarine, due to the mandatory vitamin D fortification of margarine in

Belgium. Food fortification is widely used in many industrialised countries for increasing

vitamin D intake (Ovesen et al, 2003). As the consumption of fatty fish was significantly

higher for the highest vitamin D tertiles, a lot of girls, and also boys, would benefit from

higher fatty fish consumption. On the other hand, it must be taken into account that advising

increased consumption of fat-rich food items to increase the vitamin D intake will

simultaneously increase the total fat intake. When considering vitamin D, it must also be

mentioned that ultraviolet-induced skin production constitutes the main contributor to

vitamin D in humans, making oral intake nonessential in principle (Sichert-Hellert et al,

2006). In contrast, Ovesen et al (2003) stated that skin synthesis of vitamin D may not

compensate for the low nutritional intake in Europe. Moreover, Lamberg-Allardt (2006)

suggested that the recommendation should be increased to be at least 10 µg/d in all age

groups when solar UVB is scarce. In this study population, even the 95th percentile does not

reach this level, with the 95th percentile of the girls being only 4.9 µg/d. Nevertheless,

Ricketts disease, the direct consequence of vitamin D deficiency, is not a major public health

problem in Belgium. But this does not mean that a long term, too low intake of vitamin D

will not create health problems.

The results of the PUFA intake of the studied Flemish adolescents were compared to other

intake data of (LC) n-6 and n-3 PUFAs for adolescents from

- Australia (1086 children; 12-18y; one 24-h recall plus a food frequency questionnaire

(FFQ); 1995) (Howe et al, 2006) and

- the USA (581 boys and 536 girls; 14.8±0.02y; one 24-h recall; 2001) (Harel et al, 2001).

The latter only described the intake of n-3 PUFAs. The LA and ∑n-6 PUFA intake was higher

in Belgium than in Australia (respectively, 11.7 and 11.8 versus 11.1 and 11.4 g/d). The LNA

intake was highest in Belgium, followed by Australia (1.18 g/d) and the USA (0.35 g/d).

Nevertheless, the sum of EPA, DPA, and DHA was highest in Australia (195.0 mg/d),

followed by Belgium (185.7 mg/d) and the USA (38.5 mg/d).


70

4.2.2.3. PUFA intakes of young women

International data of mean dietary intake of PUFA in adult women, expressed in g or

mg/day, are presented in Table II.9.

Table II.9: Fat and PUFA intake data of different other larger and smaller population studies with women (w)

Country Population Method Fat LA LNA AA EPA DPA DHA ∑ n-6

* ∑ n-3

# LA/

LNA n-6/ n-3

Larger sample g/d mg/d g/d Belgium (this study) 641 w (18-39y)

2-day record 77.6 11.6 1.4 56 78 25 131 12.0 1.7 8.7 7.8

Australia (Howe et al, 2006) 5770 w (+19y) 24h-recall NA 8.7 0.9 117 60 52 83 8.9 1.1 NA NA France (Astorg et al, 2004) 2,785 w (35-63y)

6x 24h-record 73.6 8.1 0.7 152 118 56 226 NA NA 11.1 NA

Germany (Linseisen et al, 2003) 898 w (35-64y) 24h-recall 76.1 11.6 1.5 140 80 NA 140 11.7 1.7 NA 7.2 Germany (Linseisen et al, 2003) 1,078 w (35-64y) 24h-recall 78.0 10.9 1.3 160 70 NA 140 11.0 1.5 NA 8.0 Australia (Meyer et al, 2003) 3178 w (19-64y)

24h-record NA 9.4 1.0 41 46 21 89 9.4 1.2 8.0 NA

Japan (Tokudome et al, 1999)

180 w (middle-aged)

1-day record 57.7 11.2 1.7 139 242 NA 469 NA NA NA NA

Norway (Johansson et al, 1998) 1627 w (16-79y) FFQ 67.0 8.8 1.2 120 270 60 400 NA NA NA 4.2

Smaller sample USA (Loosemore et al, 2004) 31 w (mean:25y)

2x 24h-recall 74.0 11.6 1.3 155 16 NA 68 11.7 1.5 9.0 4.5

Japan (Tokudome et al, 2003) 71 w (40-49y)

7-day record NA 10.5 1.6 135 278 75 497 10.6 2.4 NA 4.6

Canada (Innis & Elias, 2003)

55 pregnant w (20-40y) FFQ 79.8 11.2 1.6 121 78 NA 160 NA NA NA NA

Belgium (De Vriese et al, 2001)

26 pregnant w; 1st trim FFQ 85.9 12.9 1.3 130 170 NA 300 13.2 1.8 NA NA

Belgium (De Vriese et al, 2001)

26 pregnant w; 3rd trim FFQ 90.2 13.7 1.5 130 150 NA 300 14.0 2.0 NA NA

Nether- lands (Otto et al, 2001)

19 pre-pregnant w (mean:30.7y) FFQ 88.2 13.2 1.1 30 50 10 90 13.2 1.3 NA 11.3

Nether- lands (Otto et al, 2001)

20 pregnant w (mean:30.7y) FFQ 86.9 13.2 1.0 20 80 20 140 13.2 1.3 NA 12.0

∑n-6 * and ∑n-3 # = depending on the study; NA= not available; w = women; trim = trimester

- The women of the studies in France, Norway, and Australia seemed to consume less LA,

whereas the women in Germany and Japan consumed quite similar levels as the Flemish

women (Astorg et al, 2004; Howe et al, 2006; Johansson et al, 1998; Linseisen et al, 2003;

Meyer et al, 2003; Tokudome et al, 2003). Conversely, higher absolute LA intake levels


71

were found in another Belgian and a Dutch study with a smaller sample size (De Vriese

et al, 2001; Otto et al, 2001) (Table II.9).

- Another pattern was found when comparing the LNA intake. The absolute intake of

LNA determined for the Flemish women is quite high when compared to other studies,

with exception of the Japanese and German study (Loosemore et al, 2004; Tokudome et al,

2003) and a study done in pregnant Canadian women (Innis & Elias, 2003). In all the

other countries, a lower LNA intake was determined. This is remarkable, since a

comparison of the LNA intake with the Belgian recommendations showed a deficit of

this fatty acid in the diet of the Belgian women. Astorg et al (2004) translated the mean

LNA intake of the female population group in %E, being 0.38 %E. Comparing to the

Belgian recommended minimum level, the French women have even a worse deficit than

the Flemish women. Correspondingly, mean LA/LNA ratio of the French women is

higher (11.1) than the one found for the Flemish women.

- Most of the other studies found a mean absolute AA intake, being two or three times

higher than the mean AA intake of the Flemish women. Nevertheless, the most

contributing food sources for AA are equal compared to those found in the French and

the Australian study (Astorg et al, 2004; Meyer et al, 2003), being meat, poultry & eggs

(counting for more than 50% of the AA intake) and fish & seafood. So a probable

explanation is that the AA concentration of these food items is underestimated in the

FCDBs that were used in this study.

- A large variation is found in the intake estimations of LC n-3 PUFAs between studies

and countries. As for LA and LNA mean intake, the mean LC n-3 PUFAs intake of

Flemish women is comparable with the results found in the German study (Loosemore et

al, 2004) as well as with the results of the Dutch study (Otto et al, 2001). A larger absolute

intake for EPA, DPA and DHA is found in France and especially in Norway and Japan,

probably due to larger seafood consumption. In this study, the average seafood

consumption of the women (calculated as a mean of the two days) was 27.0 g/d. Welch et

al (2002) reported the seafood consumption in 10 European countries for men and

women between 35 and 74 years old. For women, a mean seafood consumption of 15.9

g/d and 19.9 g/d was found in two German cities; and a mean of 13.3 g/d and 13.4 g/d

in two Dutch cities (Welch et al, 2002). In comparison, French women have a mean

seafood consumption ranging between 35.0 g/d and 52.4 g/d, depending to the region

where they live; and the mean intake of Norwegian women was 42.9 g/d in South & East

Norway and 63.3 g/d in North & West Norway (Welch et al, 2002). Johansson et al (1998)


72

even reported a mean intake of 57 g fish per day for Norwegian women. These results

show that the higher seafood consumption in the latter two countries is clearly the main

reason for the higher LC n-3 PUFA intake.

Comparison of food sources of nutrients over different studies is not straightforward, since

the composition of food groups are not the same in the different studies. The composition of

food groups in this study is based on the French study (Astorg et al, 2004), so the results of

these two studies are quite comparable. The comparison with other results must be seen as

an indication.

- The main sources for LA intake of Flemish women were fats & oils (31.4%), followed by

cereal products (16.4%). The important contribution of fats & oils to the LA intake was

also found in the French and the Australian studies (counting for 33.5% and 21.9%,

respectively) (Astorg et al, 2004; Meyer et al, 2003) as well as in the Japanese study

(Tokudome et al, 1999).

- Considering the LNA intake of Flemish women, fats & oils were by far the major

contributor (45%), especially fatty sauces and margarines. This is at variance with the

French study, where fats & oils were only a minor contributor (10%), the major ones

being dairy products, fruits & vegetables and meat, poultry & eggs (Astorg et al, 2004).

Fatty sauces and margarines seem to be much less consumed in France than in Flanders.

- For AA, the intake of Flemish women was mostly contributed by meat, poultry & eggs,

followed by fish & seafood. The same results were found by the French and Australian

authors: meat, poultry & eggs counted for 67.2% and 70.2%, respectively; followed by

fish & seafood (11.1% and 27.2%, respectively). Also in the Japanese study, chicken eggs

and pork together contributed for 50.6% to the AA intake.

- In this study, fish & seafood were the major contributors for the EPA, DPA, and DHA

intake counting respectively for 87.3%, 66.0%, and 80.0%. In Norway, fish was also the

major source (respectively counting for 55%, 43% and 57%), followed by cod liver oil,

which was used by 36% of the survey subjects (Johansson et al, 1998). In contrast, cod

liver oil, being rich in LC n-3 PUFA, was not consumed by the Flemish women included

in this study. In the French data, fish & seafood were the major sources for EPA and

DHA, respectively for 72.0% and 64.7% (Astorg et al, 2004), but meat, poultry & eggs

were the first contributor of DPA intake (55%). In Norway, meat and fish were the main

sources of DPA (43% each) (Johansson et al, 1998). In this study, the contribution of meat,

poultry & eggs to the DPA intake was lower (31%). A very recent Australian publication


73

described in detail the contribution of meat sources to the intake of LC n-3 PUFA (Howe

et al, 2006). Their analyses showed that fish & seafood contributed for 49.7%, 15.4%, and

69.9% to the intake of respectively EPA, DPA, and DHA. On the other hand, meat,

poultry & game products counted for respectively, 44.8%, 73.2%, and 19.6%. This shows

that an underestimation of the LC n-3 PUFA content of meat can be crucial, since the

consumption of meat & meat products is large compared to that of fish & seafood.

5. Conclusion

The intake of LA for the studied Flemish population groups fell within the recommended

range. In contrast the intake of LNA was rather low and as a result, the mean LA/LNA ratio

was high. Therefore, the Flemish population would benefit from a higher consumption of

LNA rich foods e.g. by replacement of n-6 rich oils by n-3 rich oils such as linseed and

rapeseed in food formulations.

Moreover, dietary shifts are necessary to bridge the gap between the intakes and the

recommendations of LC n-6 and n-3 PUFAs. Regular replacement of meat products rich in

SFA by poultry meat is a possible solution to increase the AA intake, which was evaluated to

be very low for the studied pre-school children. Fatty fish consumption should be stimulated

since it is a rich source of LC n-3 PUFAs of which the current intakes are far below the

recommendations. Seafood consumption data of the overall Belgian adult population

indicated that low intakes of LC n-3 PUFAs are omnipresent. Increased seafood consumption

can, additionally, lead to higher vitamin D intake and can replace SFA-rich food items.

Nevertheless, due to contamination, seafood consumption can not be increased without

limitation. In chapter IV of this thesis it was studied how much seafood can be consumed per

week without leading to contaminant intake of toxicological concern.

However, further investigation is necessary to explore the best way to convince the Flemish

population of these health beneficial steps. Since a previous study in Belgium showed that

consumer awareness and beliefs related to n-3 PUFA, their health benefits and their food

source are poor and often wrong (Verbeke et al, 2005), nutrition education should play an

important role in order to convince people of this dietary shift.

Chapter III. Elaboration of databases

75

Chapter III.

Traceability and nutrient and contaminant content of seafood on the Belgian market: elaboration of databases


1. Sioen I, Verbeke W, De Henauw S, Parmentier K, Raemaekers M, Willems JL, Van Camp

J. Traceability of seafood products on the Belgian market. Fisheries Research. Submitted.

2. Sioen I, De Henauw S, Verdonck F, Van Thuyne N, Van Camp J. Development of a

nutrient database and distributions for use in a probabilistic risk-benefit analysis of

human seafood consumption. Journal of Food Composition and Analysis 2007; 20(8): 662-670.

3. Sioen I, Van Camp J, Verdonck F, Van Thuyne N, Willems JL, De Henauw S. How to use

secondary data on seafood contamination for probabilistic exposure assessment

purposes? Main problems and potential solutions. Human and Ecological Risk Assessment

2007; 13(3):632-657.


76

1. Introduction

One of the main objectives of this PhD-thesis was to execute an intake assessment of

nutrients and contaminants via seafood consumption in order to evaluate the risks and

benefits related to seafood consumption in Belgium and to formulate recommendations

concerning seafood consumption. This intake assessment is based on combining seafood

consumption data with nutrient and contaminant concentrations – if possible taking into

account the origin of the seafood species – by using a probabilistic approach, which takes

into account the existing variability of the different parameters (consumption data, body

weight data, and concentration data). Taking into account the variability of parameters is

relevant in the case of contaminant and nutrient concentrations in seafood as well as for

consumption patterns and body weights. In the probabilistic intake assessment executed in

this PhD-thesis and described in chapter IV, the variability of the seafood consumption by

the population is taken into account in a non-parametric way (i.e. using all the individual

data as such and without assuming any underlying probability model), whereas the

variability of the nutrient and contaminant concentrations is taken into account in a

parametric way (i.e. by applying and assuming probability distributions). Moreover, special

attention was given to the origin and traceability of the seafood consumed in Belgium.

Nutrient and contaminant concentration data can be collected:

1. by performing laboratory analyses in representative samples, or

2. by using published literature data (articles, reports, food composition tables, etc.).

In this PhD-thesis, the second approach – retrieving literature data - is applied. This is in

accordance with the strategy recently proposed by Brüders et al (2005), stating that existing

data should be used in the most effective way as collecting samples and analyzing them is

expensive. Nevertheless, the comparison of data from different sources is always difficult

especially if one only has access to aggregated data without having sufficient information

about the collecting scheme, location data, sampling procedure, or laboratory procedures

(Brüders et al, 2005). Moreover, necessary data are sometimes lacking, which hampers the

intake assessment.


77

Different databases were constructed:

1. a first database attempting to describe the origin of the seafood products available on the

Belgian market (described in part 2 of this chapter),

2. a second database describing the nutrient concentrations of seafood species relevant for

Belgian consumption (described in part 3 of this chapter), and

3. a third one describing the contaminant concentrations in the relevant species (described

in part 4 of this chapter).

This chapter starts by presenting an overview of the species that were relevant for Belgian

consumption and are as such considered in this study. On the basis of the food consumption

databases used further on in this PhD-thesis, 41 seafood species and two seafood products

(caviar and surimi) were considered. Five fat groups (FG) were determined in order to group

the species according to their fat concentration (expressed on fresh weight basis) (Table III.1).

Table III.1: The 41 seafood species and two seafood products taken into account because of their relevance for consumption by the subgroups of the population studied

Fat Group 1: < 1.0% fat

Fat Group 2: 1.0% ≥ fat < 2.5%

Fat Group 3: 2.5% ≥ fat < 5.0%

Fat Group 4: 5.0% ≥ fat < 10%

Fat Group 5: fat ≥ 10%

Anglerfish Common shrimp Anchovy Milkfish Eel Brill Common whelk Caviar Sardine European catfish Cod European plaice Conger Trout Herring Crab John dory Halibut Mackerel

Haddock Lobster Sea bream Salmon Ling Mussel Swordfish Sprat

Saithe&Pollack Nile perch Tuna Scallop Norway lobster Wolf fish Skate Oyster

Surimi Redfish Whiting Scampi

Sole Squid Tilapia Turbot

In Annex III.1 of this chapter, a table is given with the English, Dutch, French, and scientific

names of the 41 seafood species.


78

2. Traceability of seafood products on the Belgian market

2.1. Introduction on traceability

Contaminant concentrations in seafood species depend strongly on the species itself, its

metabolism, its feed, and the environmental conditions, i.e. chemical contamination of the

sediment and suspended particulate matter of the region(s) where it was living before its

catch or during its production (Domingo et al, 2007a; Judd et al, 2003a). Since Belgium is a

small country with a limited coastal region and only three commercial harbours, a large part

of the seafood available on the Belgian market is imported from other countries. Moreover,

the current seafood supply and food supply in general is characterized by an increasing

globalization, with increasing import and export, hampering adequate food traceability.

Specifically, seafood is the most traded of all food commodities (Nierentz, 2006). Hence, it is

a substantial challenge to improve the seafood traceability. Traceability is defined as ‘the

ability to trace the history, application, or location of what is under study’ (ISO 9000)

(Frederiksen & Gram, 2003; Thompson et al, 2005) and is part of any good quality

management system. The Commission of the Codex Alimentarius defined traceability or

product tracing as ‘the ability to follow the movement of a food through specified stage(s) of

production, processing, and distribution’ (FAO/WHO, 2004).

The purpose of this traceability study was to trace the origin of seafood products available

on the Belgian market, in order to link them with the corresponding contaminant

concentrations. As such, one may regard a traceability study as a part of an epidemiological

investigation (Frederiksen & Gram, 2003). The introduction of traceability into the food

supply chain is a relatively new concept, whereas traceability systems have been used for

many years in several other sectors such as aviation, automobile, and pharmaceutical

industry (Frederiksen & Gram, 2003). Globalization of trade and the lack of international

standards made it difficult to identifying the origin and history of seafood products, raising

concerns from retail, food services, and consumers about the safety of their seafood supplies.

It is clear that traceability could be an important strategy to address consumer concerns

about quality of the supplied seafood and declining fish populations as well as to address

growing pressure from consumers to produce sustainable food (Thompson et al, 2005).


79

In 2001, the European Commission published a regulation laying down detailed rules about

the information that has to be supplied to the consumer of fishery and aquaculture products

(Commission Regulation EC 2065/2001) (European Commission, 2001). This regulation

determines that appropriate marking or labelling for seafood products has to indicate:

1. the commercial designation of the species,

2. the production method (caught at sea or in inland waters or farmed), and

3. the fishing ground where it was caught or produced (i.e. traceability).

The latter must be documented as follows:

- for fish caught at sea, the FAO area (for more details see later) must be stated;

- for fish from inland waters the country of origin must be given;

- for farmed fish the country of the final development of the product must be given

(Frederiksen & Gram, 2003).

This information must be indicated on the label or posted up in the case of fresh fish sold in

bulk, e.g. in retail or fish shops.

Up to now, no database is available describing quantitatively the origin of the seafood

species available for consumption in Belgium; and to the authors’ knowledge no publications

describe the existence of such a database in other countries. Moreover, a pan-European study

by Euroconsumers showed that incorrect labelling of seafood products is the rule rather than

the exception: almost 90% of the samples were labelled incorrectly (Jooken & Lauryssen,

2006). Also a recent Norwegian study evaluated the traceability systems in the supply chain

of the Norwegian fish industry and food retail trade and showed that traceability labelling is

unsatisfactory (Karlsen & Senneset, 2006).

The aim of the work reported in this part of chapter III was to investigate the traceability and

origin of the seafood products available on the Belgian market, with the purpose of using

this information for assessing contaminant intake of the Belgian population via seafood

consumption. However, during the study, it became clear that many impediments exist

which made the detailed execution of this work difficult. The text, therefore, focuses on the

attempts made, the results obtained, and recommendations related to extra data needed to

improve seafood traceability and any subsequent analyses related to benefits and/or risks

depending on seafood origin. Probably, the origin of the seafood products also influence the

nutrient content, but the information on nutrient concentrations was not detailed enough to

take this into account in the same way as was done for the contaminant concentrations.


80

2.2. Materials and methods

In order to gather information about the origin of the 41 commercial seafood species

available on the Belgian market (Table III.1 given in the introduction part of chapter III), four

different data sources were combined:

Two national databases:

1. An economic database from the Central Economic Council (CEC) being part of the

Ministry of Economic Affairs of Belgium (received in November, 2004);

2. Data on landings in Belgian harbours, provided by the Sea Fisheries Department,

Agricultural Research Centre (received in November, 2004);

And two international databases:

3. The landings/production databases of seafood from the Food and Agriculture

Organisation (FAO) (together with the software FishStat Plus Version 2.3)

(http://www.fao.org; consulted in January 2005);

4. Catch data from the International Council for the Exploration of the Sea (ICES)

(http://www.ices.dk; consulted in January 2005).

The economic CEC-database resulted in an Excel-file containing Belgian import and export

data of all seafood products (fresh, frozen, canned). The database contained for each seafood

product the flow (import or export), the product name (mentioning one or more scientific

names of seafood species), the Dutch name, the country of import or the country to which it

was exported, the amount in tons imported or exported, and the value in euros. The FAO

provided data about the annual catch and production data of all different seafood species by

all countries. ICES collected data on annual landings officially submitted by 19 ICES

Member States in the Northeast Atlantic Sea including over 200 species. For this study, a

combination was made of different FAO-datasets and the ICES-dataset in order to link this

combined datasets with the CEC-data and describe as detailed as possible the catching and

production of seafood with respect to the various fishing grounds.

The FAO classification was applied to describe the origin of the seafood species. The FAO

defined worldwide 24 different fishing grounds. Six codes describe a zone of inland waters

and 18 describe a sea or a part of an ocean (Table III.2 and Fig. III.1).

http://www.ices.dk/


81

Table III.2: Area codes and names of the 24 international fishing grounds all over the world (www.fao.org) Area code Area name

1 Africa - Inland waters 2 America, North - Inland waters 3 America, South - Inland waters 4 Asia - Inland waters 5 Europe - Inland waters C

ontin

ents

6 Oceania - Inland waters

21 Atlantic, Northwest 27 Atlantic, Northeast 31 Atlantic, Western Central 34 Atlantic, Eastern Central 37 Mediterranean and Black Sea 41 Atlantic, Southwest 47 Atlantic, Southeast 48 Atlantic, Antarctic 51 Indian Ocean, Western 57 Indian Ocean, Eastern 58 Indian Ocean, Antarctic 61 Pacific, Northwest 67 Pacific, Northeast 71 Pacific, Western Central 77 Pacific, Eastern Central 81 Pacific, Southwest 87 Pacific, Southeast

Part

s of

an

ocea

n

88 Pacific, Antarctic

Fig. III.1 The 24 international fishing grounds as defined by FAO (www.fao.org)


82

In addition, it was of interest to subdivide four fishing grounds and define the origin by their

subdivisions, since they are important regions for the seafood on the Belgian market:

1. the North-eastern Atlantic Ocean,

2. the Eastern Central Atlantic Ocean,

3. the South-eastern Atlantic Ocean,

4. the Mediterranean & Black Sea.

As such, smaller seas, e.g. the Baltic Sea and the North Sea, became separated entities and

were not just considered together with a lot of other seas as the North-eastern Atlantic

Ocean.

The applied methodology to determine the origin of the available seafood products

consisted of two consecutive steps:

1. In a first step, the countries of origin were determined for the seafood products on the

Belgian market.

2. In a second step, an attempt was made to express their origin in terms of fishing

grounds.

2.2.1. STEP 1: Defining the countries of origin

With respect to seafood imports, only the import data from other countries into Belgium

were taken into account in this phase (step 1) of the data management process. Thus, Belgian

landing data (i.e. catching data of the Belgian fleet) were excluded at this stage, since these

data were provided by the Belgian Sea Fisheries Department already containing details

about the origin (expressed on the basis of the FAO classification). Moreover, only data of the

year 2000 were used, describing quantitatively the different countries of import for each

seafood product on the Belgian market in 2000.

Considering the seafood imported from other countries to Belgium, a large amount of the

imported seafood is again exported, which need to be taken into account in step 1. Lacking

more detailed information, the assumption had to be made that exporting seafood products

from Belgium to other countries (whether it were own landings or imports) did not change

the country’s proportional import share. The result of this first step was that for each seafood

product available on the Belgian market the ratio coming from each country of interest was

calculated in terms of percentages.


83

2.2.2. STEP 2: Defining the fishing grounds of origin

In this second step, the relative amounts per country were converted to relative amounts per

fishing ground by applying the combined dataset of the FAO data and the ICES data. To do

so, the data were sorted per species (based on the scientific name) and per country (result of

step 1) and for each of these species-country combinations the amounts caught or produced

per fishing ground were given in tons (result of step 2). In step 2, a second assumption had

to be made, considering that the country of import is the same as the country of origin

(assuming that there was no transit).

In practice, two databases were constructed: CEC2000 and FAO2000. CEC2000 describes the

amount imported from all relevant countries (species-country combination), for each seafood

product on the Belgian market (defined by a product name). FAO2000 describes the amount

caught or produced in each relevant fishing ground (species-country-fishing ground

combination), for each seafood species and all countries all over the world. These two

databases were then linked to each other at the level of species and country by creating 101

Sp-codes (Species-codes) and 1022 unique SpC-codes (Species-Country-codes). As such, the

species-country combinations in both files could be described by that code. Subsequently, the

distribution per SpC-code out of CEC2000 over the different fishing grounds was calculated

case by case by multiplying the amount imported in ton by the relative percentage caught or

produced by that country over the different fishing grounds (found by looking up the

corresponding SpC-code in FAO2000).

Up to this point in the procedure, the data describing the landings in the Belgian harbours in

2000 were not taken into account. These data were available in a database already split up

per fishing ground. Subsequently, these data were added to the newly composed database.

Thereupon, for each Sp-code the distributions over the fishing grounds were summed in

order to get the overall division per species without the separation per country. Finally, the

relative percentages for each species per fishing ground were calculated to reach the

objective of the study.


84

2.3. Findings

2.3.1. Countries of origin (result of step 1)

From the CEC2000 data, it appeared that in the year 2000 219,000 tons of seafood was

imported from 116 countries to Belgium, spread over the five continents and 26,000 tons

seafood was landed in Belgian harbours. This yields a total amount of 245,000 tons of

seafood entering Belgium in 2000 from which 89% was imported. Of the total of Belgian

imports and landings, 90% was obtained from only 22 countries (Belgium inclusive) (Table

III.3), 71% of which originated from European countries. When different seafood groups are

considered, it appeared that 98% of the shellfish was imported, and imported finfish

accounted for 85% of the total finfish supply. As shown in Table III.3, more than 50% is

supplied by Belgium and three European countries: The Netherlands, France, and Denmark.

It is important to indicate that the table describes the countries of import, not of origin. Of

the total amount of 245,000 tons of seafood entering in Belgium, 40% (99,000 tons) were

subsequently exported to other countries, leading to 146,000 tons available on the Belgian

market for consumption. This is roughly 14.6 kg/year/caput or 280 g/week/caput.

Table III.3: The 22 most important countries supplying seafood for the Belgian market, with their percentage supplied relative to the total amount (% of 245,000 tons) The Netherlands 23.9 Vietnam 2.0 Thailand 1.1 Belgium 10.6 China 1.9 Senegal 1.0 France 9.3 India 1.7 Ireland 1.0 Denmark 7.7 Sweden 1.6 Uganda 0.9 Germany 7.0 United States of America 1.6 Indonesia 0.9 Tanzania 6.4 Canada 1.5 Ecuador 0.9 United Kingdom 3.7 Spain 1.5 Iceland 2.4 Bangladesh 1.3

2.3.2. Fishing grounds of origin (result of step 2)

Two assumptions had to be made before distributing the imported seafood products on the

Belgian market over the different fishing grounds expressing their origin.

1. The first one involved assuming that the country of import did not differ from the

country where the seafood was captured or produced. It should be noted that this is not

correct in all cases since it is known that several countries import raw fish from a country

with an extensive seafood capture or production capacity to process them (peeling of

shrimps, filleting of fish …) and export post processing. But as no quantitative

information was available concerning transit of seafood products, there was no

alternative for this assumption.


85

2. A second assumption was related to the observation that a large amount of fish imported

or caught in Belgium is exported once again to other countries. This hampers the

determination of the origin of marine food products, since it is unknown what the origin

of the exported products really is: they may be landed or imported from various

countries. The assumption was made that exporting a certain seafood species out of

Belgium to other countries (whether it were own landings or imports) did not change the

ratio describing the different countries of origin for that species.

Within the assumptions made, the combination of the estimated data indicated that more

than 50% of the seafood products on the Belgian market originated from the Northeast

Atlantic Area, with the North Sea being the most important sub area (accounting for 13%).

Fig. III.2 shows the results for two individual seafood species frequently consumed in

Belgium. This figure indicates that almost all cod available on the Belgian market origins

from the Northeast Atlantic Sea. Tuna originates from many different FAO areas, with the

Eastern Central Atlantic Sea contributing the most.

Cod

0%

5%

10%

15%

20%

25%

30%

35%

40%

21 27_01 27_02 27_03A 27_03B_C 27_03D 27_04 27_05 27_07 27ns

Tuna

0%

5%

10%

15%

20%

25%

30%

35%

21 31 34 37 41 47 51 57 61 67 71 77 81 87 27_08 27_09 27ns

Fig. III.2 The calculated origin of two frequently consumed fresh fish species in Belgium (in %)

Area code Area name (FAO) 21 Atlantic, Northwest 27 Atlantic, Northeast 27_01 Barents Sea 27_02 Norwegian Sea, Spitsbergen

& Bear Island 27_03A Skagerrak and Kattegat 27_03B_C Sont and Belten 27_03D Baltic Sea 27_04 North Sea 27_05 Iceland and Faroer Islands 27_07 Irish Sea and coast 27_08 Gulf of Biskaje 27_09 Portuguese coast 31 Atlantic, Western Central 34 Atlantic, Eastern Central 37 Mediterranean and Black Sea 41 Atlantic, Southwest 47 Atlantic, Southeast 51 Indian Ocean, Western 57 Indian Ocean, Eastern 61 Pacific, Northwest 67 Pacific, Northeast 71 Pacific, Western Central 77 Pacific, Eastern Central 81 Pacific, Southwest 87 Pacific, Southeast


86

2.4. Discussion and conclusion

Pàlsson et al (2000) reported that the fish industry mainly uses paper based traceability

systems. Although this is gradually changing owing to information technology

advancements (Cibot & Kolypczuk, 2003), important barriers to its application in the field

persist. Software systems allowing for the integration of financial and production data in one

software package are typically too costly for the small business units in the fish industry

(Frederiksen & Gram, 2003). Questions rise as to how artisanal and small exploitations at the

primary production level can cope with regulatory and contractual traceability requirements

(Lupin, 2006). Even when these practical issues would be solved, the lack of an electronic

database bringing together all relevant information on (inter)national level remains a major

challenge related to the traceability of commercial seafood. This lack of one database causes

problems to determine the origin of the products and might also hamper withdrawal or

recall of defective or hazardous products. Moreover, even if the data resulting from

European legislation had led to a consistent database, the enforced level of detail might not

be sufficient, e.g. the Baltic Sea cannot be distinguished from the North Sea, since both

fishing grounds are located in the FAO area North-eastern Atlantic Ocean.

Lupin (2006) notes that ‘despite the noticeable development of traceability systems, some

important questions remain open, particularly at the level of international food and fish

trade’. Apart from regulatory issues, approvals of principles and practical implications, one

of these open questions relates to the establishment of centralised databases. This traceability

study made clear that it is very hard – even in the case of a small and well-defined market –

to trace the origin of commercial seafood products until consumers’ dishes; this, in spite of

all existing regulations concerning traceability of food items and marking and labelling of

commercial food products. Examples of voluntary chain traceability systems including

continuous recording and maintaining of all relevant information on a central database are

available (e.g. TRACEFISH). Nevertheless, to the authors’ knowledge, no existing market-

level databases covering fish species, volumes, and origin are available. In absence of a clear

and useable database, this study aimed at establishing the origin of seafood products on the

Belgian market. Therefore, efforts were made to collect data from different sources and to

link them.


87

Even though the market under consideration is rather small and well defined, important

problems were encountered. First, the information needed came from different, non related

sources, which can hardly be linked. Second, countries of import did not necessarily define

fishing grounds or product sites and several assumptions had to be made in this respect.

Third, even when one is able to retrieve information about the origin of seafood in terms of

fishing grounds, the relevance of this information remains uncertain since seafood species

caught at a certain fishing ground might have transited many others during their life.

Finally, the purpose of studying the origin of seafood products was to use the numerical

results in the assessment of the intake of contaminants via seafood consumption. This also

necessitates the availability of contaminant concentration data in seafood products

specifying the related fishing grounds. In contrast to contaminants, for nutrient

concentration in seafood not enough detailed information was available specifying the

related fishing grounds. Therefore, for nutrients the origin was not taken into account.

It is indeed a challenge to investigate the traceability of seafood on national markets, since

seafood is the most traded food item and given its substantial health benefits and eventual

safety risks for consumers. Moreover, since regulations exist on European level to trace and

label seafood with its origin, some extra efforts on the level of data collection can hopefully

lead to (inter)national databases making seafood much more traceable. When the latter

efforts are performed, it will be of interest to label the origin of the seafood on such a level of

detail that it becomes more meaningful for consumers, policy makers, and researchers in the

field of public health. This exercise has exemplified some of the difficulties encountered in

determining the origin of seafood products on a particular market and aimed at identifying

limitations and questions, which hopefully can be solved when traceability systems become

a truly operative reality after some transitory period (FAO, 2004).


88

3. Development of a nutrient database for use in a probabilistic risk-benefit analysis of human seafood consumption

3.1. Introduction

The probabilistic approach for the intake assessment of contaminants in foods is increasingly

being used (Carrington & Bolger, 2002; Gibney & van der Voet, 2003; Gilsenan et al, 2003;

Lopez et al, 2003; Matthys et al, 2005; Solomon et al, 2000; Tran et al, 2004; Vrijens et al, 2002;

Wenning, 2002). In contrast, for the assessment of the nutrient intake it is still rather

uncommon (Rubingh et al, 2003). However, since many foods show considerable variability

in nutrient concentrations, this approach is for many purposes more informative than the

deterministic approach. Particularly for seafood, the fat concentration and the fatty acid

composition is highly variable and largely depending on environmental factors like

temperature of the water, season, trophic conditions in various habitats, feed, etc. (Copeman

& Parrish, 2004; Mattila et al, 1997). By using probability distributions to describe the fat

concentration, the fatty acid composition and other nutrients, this environmental variability

will maximally be incorporated into the intake assessment. Moreover, as about 89% of the

seafood products on the Belgian market are imported from a range of countries (part 2 of this

chapter), it seems useful to take into account data from different national food composition

databases (FCDBs).

A key component in a probabilistic intake assessment is the selection of the most appropriate

probability distributions for any given data set, whether it be nutrients, contaminants, or

other parameters (Gilsenan et al, 2003). It is recognized that the determination of such

distributions will have to rely on the collection of quantitative data being accurate (i.e.

agreeing with the actual concentration) and representative (i.e. reflecting the concentration of

the whole group, in this case the species in general). This chapter part describes the problems

encountered when nutrient concentration data for seafood were brought together in a new

extensive database in order to fit useful probability distributions. The next part of this

chapter (part 4 of chapter III) handles the same topics related to the contaminant

concentration data.


89


An Excel®-database containing nutrient concentrations for seafood species was build in order

to have a dataset for the distribution fitting. Total fat content, EPA plus DHA concentration,

vitamin D, and iodine have been taken into consideration. The following sources were used:

- fourteen FCDBs (Carnovale & Marletta, 2000; Danish Institute for Food and Veterinary

Research, 2005; Favier et al, 1995; Food Standards Agency, 2002; Health Canada, 2005;

Holland et al, 1993; Institut Paul Lambin, 2004; National Public Health Institute of

Finland, 2004; NEVO Foundation, 2001; NUBEL, 1999; Salvini et al, 1998; Souci et al, 2000;

Sugiyama Jogakuen University, 2004; US Department of Agriculture and Agricultural

Research Service, 2005),

- 44 peer reviewed research papers (mostly about original research),

- two books (Ackman, 2000; Sondergaard & Leerbeck, 1984),

- own analytical data.

The selection of the FCDBs is based on their availability, either as a book or free available via

internet (found via LanguaL-website maintained and supported by the EuroFIR Consortium

and funded under the EU 6th Framework Food Quality and Safety Programme). In the new

compiled nutrient database, all relevant information was included: commercial name,

scientific name, culinary processing procedure (if relevant), farmed or wild fish (if known),

number of samples, number of individual sample units per sample in the case of composite

samples, (mean) fat content of the fish, and (mean) nutrient content, with extra statistical

data if available (standard deviation, minimum and maximum). In contrast to the

contaminant database, no distinction was made according to fishing grounds because not

enough detailed information was available to do so.

In order to perform a probabilistic intake assessment of nutrients (and contaminants) via

seafood for the Belgian populations, nutrient concentration data for all considered seafood

species were needed and distributions to describe those concentrations had to be selected.

The distribution fitting program BestFit® Version 4.5 (Palisade Corporation, Newfield, NY,

USA, 2002) was used to determine the distribution model and parameters that describe the

nutrient concentrations as well as possible. For each specie and nutrient, the concentration

data points of the new developed database were entered in BestFit® in combination with a

weighing factor, expressing the cumulative probability of occurrence for each concentration.

To do so, BestFit® needs at least three data pairs. The program identifies the best describing


90

distributions to a given dataset using goodness-of-fit tests (the method of the least squares).

The final selection of the distribution model was based on this test as well as on visual

evaluation of the probability plot (Palisade Corporation, 2002). Distribution models that can

give rise to unrealistic high or low concentrations were truncated as follows: on half of the

all-time lowest concentration at the lower end of the distribution and on the double of the

all-time highest concentration at the higher end of the distribution.

3.3. Results and discussion

3.3.1. Building up a database

During the development of the database, a number of problems with respect to

methodological issues were encountered and are presented here, along with the proposed

procedures to solve them.

A first problem was related to the description of the food items, in this case the seafood

species. From country to country, inconsistency exists in naming fish and other seafood.

Seafood can be sold in different regions or countries under the same name, but being entirely

different species, even from a different genus, and consequently having a different nutrient

composition, e.g. salmon can be used to describe Salmo salar and Oncorhynchus spp. To

further complicate matters, the opposite situation also exists, i.e. that different common

names may be applied to describe one single seafood species, e.g. dab and flounder are used

as common name to describe Pleuronectus limanda. Therefore, an unequivocal food

description is of primary importance to avoid confusion between different species when

bringing data from different sources together. In the compiled database, it was aimed to

describe all seafood species by their scientific name. In most of the publications used, the

scientific name of the species for which data were reported was given. In contrast, in some of

the applied FCDBs the food description lacked this level of detail. In these cases, an

assumption had to be made about the scientific name of the species on the basis of the

common name indicated. Based on these findings, the importance of an unequivocal food

description in FCDBs and other publications reporting nutrient composition data is once

again demonstrated.


91

A second problem was that different sampling strategies (sampling plans) were used over

the different publications reporting nutrient contents in seafood. According to Holden et al

(2002) an ideal sampling plan should address a demographic framework as well as the

profiles and attributes of the food sampled in order to obtain nationally representative

samples. When looking to the articles used:

- some researchers obtained their samples from local fish wholesalers (Mattila et al, 1995;

Sérot et al, 1998),

- others purchased them from various markets or retail stores over a certain region

(Piironen et al, 2002; Takeuchi et al, 1984),

- others tried to cover different fishing grounds (Mattila et al, 1997),

- etc.

Holden et al (2002) set up a range of criteria in order to evaluate the data according to their

sampling plan. In this study, no weighting of the data was applied according to the

representativeness of the sampling procedures since no objective, quantitative criteria were

available to do so. A difference is only made between the results of individual sample units

and composite samples. Compositing of sample units is a possible way to guarantee the

value of an individual measurement as it represents a mean of several units. It can, however,

result in the compression and underestimation of the variability (World Health

Organization, 2000). Only a minority of the articles used reports the application of composite

samples and reports the number of individual units in each sample. Nevertheless, a

distinction between both types of data (composite samples and individual sample units) was

made by weighing the data according to the number of sample units, such that

concentrations measured in composite samples get a higher probability of occurrence. A

weighing factor Wi,unit for each concentration data point xi was created. Wi,unit was equal to

one in the case of an individual sample unit and equal to the number of units per sample in

the case of composite samples. Hence, the concentration measured in a composite sample of

n individual units was considered as the mean of n individual samples (nevertheless, this

ignores the fact that n fish samples caught together are likely to have more similar

concentrations than the same number of fish samples caught separately). When the number

of individual sample units per composite sample was unknown, Wi,unit was assumed to be

one, leading to the assumption that the values from publications not giving the appropriate

details have a lower weight.


92

Third, analytical results in different publications are based on a different number of samples

and are reported in different ways. In some publications, all individual measurement results

are reported, whereas others only report mean values of a group of samples (aggregated

results). As indicated by the WHO, the preferred method for generating a distribution curve

for a nutrient or contaminant in food is to use data on individual measurements. When data

on individual measurements are not available in sufficient quantities (or quality) to generate

a distribution, the next preferable option suggested by the WHO – and adopted in this study

- is to combine data from different sources, from both individual and aggregated results

(World Health Organization, 2000). For the purpose of using the collected data in an

appropriate way as basis for the fitting of the most suitable distribution, a second weighing

factor Wi,meas was applied for each data point xi as a function of the number of measurements

on which it was based. When the number of measurements was not known, Wi,meas was

assumed to be one, thereby assigning a lower weight in the overall distribution to results

reported without detailed information. In the case of FCDBs, less than half of the FCDBs

used mentioned the number of samples used for the nutrient content determination.

According to Greenfield and Southgate (2003), the number of samples should be at least 10

for primary produce for FCDBs. Holden et al (2002) suggested even that more than 25

analytic sample units would be needed to improve the reliability of the estimated mean.

Because this is a theoretical situation and not reality, the number of samples – when not

mentioned in the FCDB - was assumed to be one. Eventually, for each data point xi in the

compiled database, the second weighing factor Wi,meas (expressing the number of

measurements) was multiplied with the first weighing factor Wi,unit (expressing the number

of individual sample units – see above), in order to have an overall weighing factor:

Wi,total = Wi,meas · Wi,unit

Subsequently, after ranking all the available data points xi per species and per contaminant,

the cumulative probability of occurrence of each data point xi was calculated as:

( )∑∑

=

n

iixF

totali,

totali,

W

W

where i is the rank number, Wi,total is the overall weighing factor, and n is

the total number of data points. Afterwards, using BestFit®, the data were administered

together with the weighing factor expressed as a cumulative probability of occurrence. An

example of a distribution showing the variability of a nutrient concentration as a function of

the probability of occurrence is given in Fig. III.3. Extra information about the influence of

using the weighing factor is given in the next part about the contaminant database.


93

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

0 5 10 15 20 25 30 35

Vitamin D concentration (µg/100 g fresh weight)

Cum

ulat

ive

prob

abili

ty

cod salmon

Fig. III.3 Cumulative probability function of the vitamin D concentration in cod and salmon

Fourthly, FCDBs report not only values from analyses done in own laboratories. Data in

FCDBs are generated:

- by chemical analyses of food samples (from the food industry, from government

agencies, …),

- by calculations (by recipes or formulations),

- by expert estimation or by copying from other FCDBs, or

- from scientific literature.

In practice:

- five of the 14 FCDBs used did so and mentioned for each value the source (Danish

Institute for Food and Veterinary Research, 2005; Health Canada, 2005; National Public

Health Institute of Finland, 2004; NEVO Foundation, 2001; Salvini et al, 1998)

- four others mentioned that some values in the database came from other sources without

reporting the specific references and without marking the copied values (Favier et al,

1995; Food Standards Agency, 2002; Holland et al, 1993; US Department of Agriculture

and Agricultural Research Service, 2005)

- five FCDBs did not give any detailed information about the sources of their data

(Carnovale & Marletta, 2000; Institut Paul Lambin, 2004; NUBEL, 1999; Souci et al, 2000;

Sugiyama Jogakuen University, 2004).

On the one hand, values were copied without any modification. But on the other hand,

values were copied from one publication to another for applying them for a similar but not

identical species or the copied values were used after recalculation. For example, the vitamin


94

D content of anchovy in the Italian FCDB (Salvini et al, 1998) was taken over as such from the

vitamin D content of sardine given in the English FCDB (Holland et al, 1993). An example of

the recalculation method is the vitamin D content of raw red salmon in the Canadian FCDB

(Health Canada, 2005), which was calculated based on the known fat content of that salmon

and the vitamin D (in µg/g fat) content of that species given in the USDA FCDB (US

Department of Agriculture and Agricultural Research Service, 2005). When detailed source

information was available, it was possible to take this information into account and then the

following rules were applied:

- when a value did not contain any new information (used for the same species without

recalculation), the value was not taken into consideration a second time;

- as soon as a copied value contained new information (used for another species or

recalculated), it was included for further use.

When no detailed information was available about the sources of the data, no distinction

could be made between copied and non-copied values and they were taken over and

considered as new values, leading to a potentially erroneous higher weighing factor.

A fifth remark concerns the application of different analytical methods for the

determination of a same nutrient over the different publications. During the last thirty years,

different analytical methodologies were applied for the determination of vitamin D. The

older fish vitamin D data are based on the application of biological methods (Sondergaard &

Leerbeck, 1984). The vitamin D contents of Sondergaard and Leerbeck (1984) based on

biological methods, varied according to the species from 0.5 to 30 µg/100 g fresh weight.

Takeuchi et al (1984) - using an HPLC method to determine the cholecalciferol (one of the

vitamin D steroids) content in fish - obtained a range from 0 to 132 µg/100 g fresh weight.

Matilla et al (1995; 1996) also applied HPLC to determine the vitamin D content in fish

species used in Finland, analysing cholecalciferol as well as 25-hydroxycholecalciferol and

ergocalciferol. The variation found over the different fresh fish species ranged from 0.3-47.7

µg/100 g fresh weight. Mattila (1995) reported that many discrepancies occurred in the

vitamin D concentrations of fish and fish products given in the English FCDB (Food

Standards Agency, 2002; Holland et al, 1993) compared to their own studies e.g. the English

FCDB indicated that cod contains only traces of vitamin D, whereas Matilla (1995) measured

6.9±0.2 µg vitamin D per 100 g raw cod. According to Mattila (1995), the English information

is based on the estimation of Holland et al (1993) that all the fish species with white flesh

contain only traces of vitamin D and this information is taken over in several other FCDBs,


95

which brings us again to the already mentioned problem of copying values. This creates a

bias when comparing the data and using them all together for fitting one distribution

expressing the nutrient concentration and its variability in a species. Moreover, this implies

another problem about how to deal with the indications ‘traces’. When FCDBs report ‘traces’,

this means that the species contains amounts of the nutrient being below the limit of

quantification (LOQ), i.e. the minimal amount necessary for quantification. A problem arises

if no numerical value of the LOQ is mentioned in the publication. In this study, this was a

frequent problem for vitamin D concentrations (17% of the data). To replace the indication

‘traces’ by real figures that can be used for the distribution fitting procedure, an assumption

had to be made about the LOQ. Salo-Väänänen et al (2000) reported LOQ of 0.01 µg/100 g for

cholecalciferol (defined as three times the detection limit (LOD), being the lowest level at

which a nutrient can be detected), whereas a report of the FSA indicates a LOD of 0.05

µg/100 g and a LOQ of 0.1 µg/100 g (defined two times the LOD) (Lawrance, 2002). Based

on that recent report, half of the LOQ (0.05 µg/100 g) is taken to replace the indications

‘traces’.

A sixth and last encountered problem concerns the nutrient data of raw versus processed

seafood products. For vitamin D concentrations, different studies showed that the effect of

processing on vitamin D is very small (Aminullah Bhuiyan et al, 1993; Lawrance, 2002;

Mattila et al, 1999; Suzuki et al, 1988). For instance, Matilla et al (1999) found that the losses of

vitamin D were smaller than 10% after household cooking. Therefore, all available vitamin D

concentration data of raw and processed seafood products are used without distinction.

Only those, describing the content in seafood products also containing other ingredients

(flower, milk, tomato sauce, …) were excluded. The same approach was applied for the

iodine concentrations.

There is quite some knowledge that processing influences the fatty acid composition of

seafood products (Agren & Hanninen, 1993; Al Saghir et al, 2004; Candela et al, 1997; Candela

et al, 1998; Gall et al, 1983; Gokoglu et al, 2004; Mai et al, 1978; Sanchez-Muniz et al, 1992; Sioen

et al, 2006b). Also own research investigated the effects of pan frying cod and salmon fillets

in margarine and olive oil on the fatty acid composition. The results showed that the effect of

pan frying of fish fillets depends on the total fatty acid content of the fish species and on the

fatty acid profile of the culinary fat used. In lean fish a significant increase in the amount of

fatty acids was detected, leading to a fatty acid profile similar to the profile of the culinary fat


96

used. In fatty fish, there was an insignificant decrease in total fatty acid content. The

alterations in the fatty acid profile of the fatty fish also changed the fatty acid profile in the

direction of that of the culinary fat. Therefore, control over the fatty acid content and

composition of the consumed food can be influenced by the selected culinary fat (Sioen et al,

2006b).

However, the existing knowledge is not extensive enough to build a database with

conversion factors for all species, all preparing methods, and all culinary fats. Moreover,

taking into account such information in the intake assessment would only be possible if this

information was also known at the consumer side. Due to this lack of knowledge, the same

approach as applied for vitamin D and iodine was also applied for the fat and EPA plus

DHA concentration data, i.e. that no distinction was made for processing. Additionally, it

was found that the variability in the EPA plus DHA concentration data of the raw samples in

the developed database was higher than the changes induced by processing, in other words,

the overall variability did not change by taking the data of processed seafood products into

account. For example, the EPA content in raw trout ranged from 100 mg/100 g fresh weight

(Hepburn et al, 1986) to 606 mg/100 g fresh weight (National Public Health Institute of

Finland, 2004). The values of the EPA content in prepared trout varied between 200 mg/100

g prepared weight (Holland et al, 1993) and 468 mg/100 g fresh weight (US Department of

Agriculture and Agricultural Research Service, 2005). Also Stolyhwo et al (2006) found that

the variability in raw herring and sprats was larger than the changes induced by smoking or

by storage.

3.3.2. Collected data and fitted distributions

The result of the collected data and their weighing factors formed the basis for the

distribution fitting and selection process. The available data are presented here as box plots.

In such a figure, each box represents the interquartile range (25th to 75th percentile). The bold

square in the box expresses the median value (50th percentile). The whiskers extend from the

boxes and indicate the upper and lower values not classified as statistical outliers or

extremes. Open circles are statistical outliers.


97

3.3.2.1. EPA & DHA

Fig. III.4 shows box plots of the published EPA plus DHA concentrations in the different

species of interest, where four or more data points were available. The species are sorted

according to their median EPA plus DHA concentration.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

Species

0

10

20

30

40

50

Con

cent

ratio

n of

EP

A&

DH

A (m

g/g

wet

wei

ght)

Fig. III.4 Concentrations of EPA plus DHA in different species, box plots. Data available from the open

literature and from national and international reports

The figure shows that trout, anchovy, herring, salmon, sardines, mackerel, and sprat contain

a high concentration of LC n-3 PUFA. Nevertheless, the figure makes clear that a high within

variability exists in the EPA plus DHA concentrations. In other words, not only the fish

species, but a lot of other factors influence the estimated PUFA concentration, leading to high

within-species variability. For three species (skate, John dory and conger) only two data

points were available and are, therefore, not shown in the box plots. Their EPA concentration

is defined by a uniform distribution, using the two data points as minimum and maximum.

For three other seafood species: Norway lobster, Nile perch, and sea bream, no EPA plus

DHA concentrations could be found. Their EPA plus DHA concentration is therefore

described by these of lobster, sole, and wolf fish, respectively, because of the similar fat

content. The distribution of the EPA plus DHA concentrations for the considered species as

well as the number of available data points per species can be found in Annex III.2 of this

chapter.

1.Sole; 2.Haddock; 3.Whiting; 4.Anglerfish; 5.Scallop; 6.Lobster; 7.European plaice; 8.Cod; 9.Scampi; 10.Crab; 11.Common shrimp; 12.Saithe and Pollack; 13.Tuna; 14.Mussel; 15.Squid; 16.Wolf fish; 17.Halibut; 18.Eel; 19.Trout; 20.Anchovy; 21.Herring; 22.Salmon; 23.Sardine; 24.Mackerel; 25 Sprat


98

3.3.2.2. Vitamin D

Fig. III.5 shows the published data of vitamin D for different species, ordered by the median

vitamin D concentration.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20Species

0

10

20

30

40

Con

cent

ratio

n of

vita

min

D (µ

g/10

0 g

wet

wei

ght)

1.Scampi; 2.Haddock; 3.Mussel; 4.Wolf fish; 5.Sea bream; 6.European plaice; 7.Saithe and Pollack; 8.Sole; 9.Cod; 10.Anchovy; 11.Tuna; 12.Whiting; 13.Halibut; 14.Mackerel; 15.Trout; 16.Sardine; 17.Caviar; 18.Salmon; 19.Herring; 20.Eel

Fig. III.5 Concentrations of vitamin D in different species, box plots. Data available from the open literature and from national and international reports

Sardine, salmon, herring, and eel are the species with a rather high vitamin D concentration,

but all with high within-species variability. Also caviar contains quite high vitamin D

concentrations. Since only few data were available for sprat (therefore, not shown in the box

plot), and since it is of the same family as herring, the distribution of herring is also used to

describe the vitamin D content of sprat. For eight non-fish seafood species (crustacean,

shellfishes and molluscs) and six fishes with a rather low fat content, the vitamin D

concentration was described by a uniform distribution with zero as minimum and the LOD

(0.1 µg/100 g) as maximum, since the available published concentration data only indicated

‘traces’. Annex III.3 shows for the relevant fish species the distribution of their vitamin D

concentration as well as the number of available data points per species.


99

3.3.2.3. Iodine

In contrast to EPA plus DHA as well as vitamin D, iodine is a water soluble nutrient.

Nevertheless, literature data showed that the iodine concentration is influenced by the fat

concentration of the fish, as well as by its habitat (saltwater fish versus freshwater fish). Fig.

III.6 shows box plots of the iodine concentration in different species.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19Species

0

200

400

600

800

1000

Con

cent

ratio

n of

iodi

ne (µ

g/10

0 g

wet

wei

ght)

1.Trout; 2.Sole; 3.Eel; 4.Sardine; 5.European plaice; 6.Herring; 7.Salmon; 8.Tuna; 9.Halibut; 10.Oyster; 11.Mackerel; 12.Red fish; 13.Tilapia; 14.Caviar; 15.Saithe and Pollack; 16.Mussels; 17.Lobster; 18.Haddock; 19.Cod

Fig. III.6 Concentration data of iodine in different species, box plots. Data available from the open literature and from national and international reports

High within species variability was found for cod and lobster. Due to lack of data for some

seafood species, data of different species with similar characteristics were aggregated and

distributions were fitted based on the aggregated data. Two times five groups were made,

related to the fat groups but separated for saltwater versus freshwater species. The

distributions as well as the number of available data points per species can be found in

Annex III.4.

3.3.2.4. Fat

The most important reason to gather data on the total fat content of the different seafood

species was to study the correlation between the intake of fat and the intake of other

nutrients and contaminants via seafood consumption. This information also formed the basis

for the determination of the different fat groups. Data about the fat content are not shown


100

here, but Annex III.5 shows the distribution of the fat concentration for the relevant seafood

species.

3.4. Conclusion

A key component in probabilistic intake assessments is the selection of the most appropriate

input distributions to describe the parameters. This part of chapter III described the

construction of a nutrient database, pooling fat, EPA plus DHA, vitamin D, and iodine

concentrations in seafood from different FCDBs and publications together with the

encountered problems related to a lack or inconsistency of information given in these

publications: food description, number of samples, sampling plan, sources of the values,

LOQ, etc. Different solutions have been proposed and the work resulted in a huge database

allowing to describe distributions of fat, EPA plus DHA, vitamin D, and iodine

concentrations and their variability in seafood species relevant for Belgian consumption. The

distribution fitting and selection procedure resulted in different distribution models for the

different species and nutrients.


101

4. Development of a contaminant database for use in a

probabilistic risk-benefit analysis of human seafood

consumption

4.1. Introduction

Different seafood contamination databases existed already (e.g. NIFES (2004); ICES

(Raemaeckers, 2005); CFSAN (US Department of health and human services and US

Environmental Protection Agency, 2004)), but none of them as such contained enough

information to describe the concentrations in all species available on the Belgian market.

Therefore, a de novo establishment of a contaminant database for use in this PhD-thesis was

envisaged. In line with the previous description of the elaboration of the nutrient database,

this part of chapter III describes the methods that were used, the problems that were

encountered, and the solutions that have been proposed in the course of

1. the development of an extensive contaminant database, using published sources only,

2. the characterization of appropriate input distributions describing the variability of

contaminant concentrations in different seafood species from different fishing grounds,

relevant for Belgian consumption patterns.


The following contaminants have been included: mercury, indicator PCBs (iPCBs), dioxin-

like PCBs (dl PCBs), PCDDs plus PCDFs (PCDD/Fs), and total TEQ-content (sum of toxicity

normalized PCDDs, PCDFs, and dl PCBs). In the database, all relevant information is

included: commercial name, scientific name, farmed or wild seafood (if known), period of

capture, age of the sampled seafood, fishing ground where the samples were caught or

produced, number of samples, number of individual sample units per sample in the case of

composite samples, mean fat content of the seafood, and mean contaminant content, with

extra statistical data if available (standard deviation, minimum and maximum).

Concentration data of dioxin-like compounds in seafood are expressed in pg TEQ/g fresh

weight, iPCB concentrations and mercury concentrations in ng/g fresh weight. In a first

phase, all available data were entered in the Excel®-database. Later, a selection was made


102

based on temporal and geographical arguments: (1) It was decided to consider only data

points from analysis in the period 1995-2005; (2) Only those species coming from fishing

grounds or countries relevant for the Belgian market were taken into account when

characterizing the variability by probability distributions.

The database search (exploring PubMed, Web of Science and Google) has retrieved 127

relevant data sources:

- 30 reports and/or databases of governments/research institutes;

- two important national data sources: a confidential database from the Belgian Food

Safety Agency, and data analysed at the Fisheries Department of Oostende, Belgium (in

the framework of an MSc thesis);

- 80 peer reviewed research papers; and

- 15 proceedings from the International Symposia on Halogenated Environmental Organic

Pollutants and POPs (Persistent Organic Pollutants).

Based on the available Belgian food consumption data, 41 seafood species and two seafood

products were retained as relevant for the average Belgian diet (Annex III.1 given at the end

of this chapter). For these species, the contaminant distributions and corresponding

parameters were examined using the BestFit® distribution fitting program Version 4.5

(Palisade Corporation, Newfield, NY, USA). In the applied statistical procedure, available

data per species were weighed for their cumulative probability of occurrence (for details, see

below). The selection of the best fit was based on a goodness-of-fit test (the method of the

least squares) as well as on visual evaluation of the probability plot (Palisade Corporation,

2002), as was done for the nutrient distributions.

4.3. Results and discussion

This section is essentially organised around five main topics that have been identified as

potential sources of bias in the process of establishing the contaminant database and

characterizing the appropriate input distributions. For each topic, it was attempted to

include:

1. a description of the problem as such,

2. a proposal for a strategy to deal with the problem,


103

3. a justification for the proposed strategy, when relevant supported with literature data,

4. tailored recommendations for future reporting of contaminant data in order to make

them more useful for other applications.

4.3.1. Building up a database

In July 2006, the seafood database contained:

- 1177 indicator PCB (iPCB, congeners 28, 52, 101, 118, 138, 153, 180) concentration data,

- 1254 dl PCB (congeners 77, 81, 126, 169, 105, 114, 118, 123, 156, 157, 167, 189)

concentration data,

- 1615 PCDD/F concentration data,

- 1139 total dioxin-like compounds concentration data, and

- 2082 mercury concentration data.

After a selection based on the seafood species/products relevant for Belgian consumption

(Annex III.1 given at the end of this chapter), 978, 1045, 1367, 982, and 1410 data were

retained, respectively.

The retrieval, storage, and characterization of contamination data from the above mentioned

literature sources was hampered by a number of very specific problems, mainly related to:

1. the sampling plan;

2. the sample handling prior to analysis;

3. the analytical methodologies applied;

4. the format of reporting the results (individual versus aggregated results); and

5. missing data.

These aspects influence the accuracy and representativeness of the data when applied for

human intake assessment. Where possible, solutions and weighing procedures were applied

to minimize the impact of these problems. The different procedures are described below.

First, different studies in the database used a variety of sampling schemes for the collection

of seafood samples. Some samples came from local markets, others from wholesalers; some

were caught by the research group while other samples originated from existing monitoring

programs in specific regions. Although the sampling plan directly affects the

representativeness of the samples, no weighing or selection of the data in relation to

differences in sampling validity has been applied, as no objective criteria exist for that


104

purpose. This approach is similar as for the nutrient database. It emphasizes the need for

standardization of sampling procedures to achieve contaminant levels which are comparable

over different studies, a need that is also expressed by the European Food Safety Authority

(EFSA, 2005a). Another problem at the level of the sampling plan, which was also

encountered for the nutrient database, involves the application of individual sample units

versus composite samples. Compositing of sample units is a possible way to assure the

representativeness of the mean, without the need to increase the number of measurements.

However, this results in underestimation of the true variability (World Health Organization,

2000). As in the nutrient dataset, a distinction between both types of data (composite samples

and individual sample units) was made by weighing the data according to the number of

sample units, such that concentrations measured in composite samples get a higher

probability of occurrence. A weighing factor Wi,unit for each concentration data point xi was

created. Wi,unit was equal to one in the case of an individual samples unit and equal to the

number of units per sample in the case of composite samples. When the number of

individual sample units per composite sample was unknown, Wi,unit was assumed to be one,

leading to the assumption that the values from publications not giving the appropriate

details, have a lower weight. This is in line with the recommendations by Judd et al (2003b)

based on the consideration that the absence of details about the methodological approaches

inevitably introduces increased uncertainty in the dataset. In the example given in Table III.4,

Wi,unit is equal to one in all cases, which means that no composite samples were used, or that

no information about it was given.


105

Table III.4: Data about the mercury concentration in sardines from two different origins, with the different weighing factors (Wi,meas, Wi,unit, Wi,total) and the cumulative probability of occurrence (F(xi)); the fitted distributions are shown in Fig. III. 12

Sardines from the Eastern Atlantic Sea *

Hg concentration (ng/g fresh weight) Wi,meas Wi,unit Wi,total ∑i

totali,W F(xi)

5 1 1 1 1 0.06 6 1 1 1 2 0.11 7 1 1 1 3 0.17

12 1 1 1 4 0.22 12 1 1 1 5 0.28 13 1 1 1 6 0.33 13 1 1 1 7 0.39 14 1 1 1 8 0.44 18 1 1 1 9 0.50 19 1 1 1 10 0.56 23 1 1 1 11 0.61 30 1 1 1 12 0.67 38 1 1 1 13 0.72 40 1 1 1 14 0.78 47 1 1 1 15 0.83 51 1 1 1 16 0.89 88 1 1 1 17 0.94

104 1 1 1 18 1.00

∑n

totali,W = 18

Sardines from the Mediterranean Sea $

Hg concentration (ng/g fresh weight) Wi,meas Wi,unit Wi,total ∑i

totali,W F(xi)

63.6 11 1 11 11 0.03 75.7 24 1 24 35 0.10 130 300 1 300 335 0.94 142 10 1 10 345 0.96 156 3 1 3 348 0.97 208 10 1 10 358 1.00

∑n

totali,W = 358

* The data about the mercury concentration in sardines from the Eastern Atlantic Sea origin from one

publication (Knowles et al, 2003). $ The data about the mercury concentration in sardines from the Mediterranean Sea origin from four

different publications (Juresa & Blanusa, 2003; Plessi et al, 2001; Sanzo et al, 2001; Storelli et al, 2003a).

Secondly, the tissue selected for analysis and the preparation methods influence the value of

the concentration data for human exposure assessment. Proper and comparable handling of

samples prior to analysis is critical (Holden et al, 2002). In the final database, only

concentrations measured in muscle tissues of fish are used, since fish liver (often used for


106

ecological measurements) is rarely consumed in Belgium. A similar problem is that, in some

studies, the skin of the fish was removed prior to analysis whilst in others it was left in the

sample. Some studies do not mention this detail at all. However, it has been shown that this

difference in manipulation of the sample may affect the contaminant concentration (Bayen et

al, 2005; Hamre et al, 2003; Salama et al, 1998). For the current database, no data selection was

made in this regard. The rationale for this was that, by including both kinds of data (with

and without skin) the variability on consumption level is covered to some extent, as some

consumers will remove the skin before consumption, while others will not. Overall, these

very specific sources of variability at the level of the sample handling, again point at the

importance of a maximally detailed description of the preparation of samples. This was also

recently emphasized by Törnkvist et al (2005), summarizing the importance of specific and

clear instructions for sample preparation when analyzing organochlorine compounds in

seafood. Their research also addressed the need to have clear regulations and to carefully

describe the seafood sample preparation procedure.

Third, different analytical methodologies can be used for determining the same substance.

In the case of (dl) PCBs, some measured individual congeners, whereas others quantified

Aroclor mixtures, i.e. commercial mixtures of different PCB congeners. It is under the form

of Aroclor mixtures that PCBs were previously marketed, e.g. Aroclor 1016, 1260, 1254, etc.

When the PCB level was quantified as Aroclor mixtures, it was sometimes difficult to

estimate the contribution of the individual PCB congeners. These different approaches in the

quantification of PCBs influence the comparability of the concentrations reported in different

publications. For example, it is difficult to compare the data of the salmon study of Hites et al

(2004a) to the salmon contamination data reported by Easton et al (2002), because Hites et al

(2004) measured Aroclor 1254 whereas Easton et al (2002) measured 112 unspecified PCB

congeners. Aroclor concentrations were not included in the database, unless it was possible

to convert them to concentrations of the seven indicator PCBs, e.g. with the data published

by Kodavanti et al (2001). In order to avoid the creation of biases, only the concentration data

of publications clearly mentioning that the seven indicator PCBs were measured (iPCBs,

congeners 28, 52, 101, 118, 138, 153, 180; (EFSA, 2005c)) and/or the sum of the four non-ortho

and eight mono-ortho dl PCBs (congeners 77, 81, 126, 169, 105, 114, 118, 123, 156, 157, 167,

189; (EFSA, 2005c)) were used in the newly developed database and applied when

characterizing the probability distributions. At the level of dioxin-like compounds, only data

expressed as pg TEQ/g fresh weight were used (where possible, concentration data


107

expressed in g/g wet weight were converted to TEQs using the TEF-values). A problem

directly related to the reading of the results from the analysis is the statistical treatment of

concentrations below the limit of detection or quantification (LOD or LOQ), a problem that

was also encountered for the nutrient database. Nevertheless, this did not cause many

problems in our study because the published data for the sum of iPCBs, sum of dl PCBs,

PCDD/F, and total sum of dl-like compounds were all present at detectable levels. The

exception is for some (n < 10) of the mercury concentrations. These have been systematically

replaced by the LOD reported in the considered publication (as worst case scenario

approach). The effect of this replacement is not explicitly studied, but assumed to be

negligible as it was applied to so few cases. On the other hand, it is likely that different

approaches have been applied in different publications in the calculation of aggregated data.

To give an example: when some of the PCB congeners were lower than the LOD during the

determination of the sum of the seven iPCBs, one author may have replaced them by zero to

calculate the reported sum of PCB congeners, while another author may have replaced it by

the LOD before adding up. Obviously, such particularities regarding the data are hidden

behind the aggregated figures in the published reports and cannot be dealt with in the

process of compiling new databases. It is nevertheless useful to consider the potential effect

of this phenomenon. This has been explicitly studied by Storelli et al (2003b) investigating the

effect of replacing the non-detectable congener concentrations by respectively zero, half of

the LOD, and the LOD, in the calculation of the dl PCB intake. By assuming a seafood

consumption of 60 g seafood per person and per day, an intake of 0.01-6.20 pg TEQ/kg body

weight (bw)/day, 0.29-6.23 pg TEQ/kg bw/day, and 0.56-6.26 pg TEQ/kg bw/day,

respectively, was obtained. This illustrates that the approach used leads only to a slight

difference in the higher percentiles of the intake distribution. It can be assumed that similar

effects apply to other aggregated data in this context.

Fourth, as also encountered for the nutrient database, analytical results in literature are

reported in many different ways. Again, a second weighing factor Wi,meas was applied for

each data point xi as a function of the number of measurements on which it was based. When

the number of measurements was not known, Wi,meas was assumed to be one. As explained

for the nutrient database, for each data point xi in the compiled database, the second

weighing factor Wi,meas was multiplied with the first weighing factor Wi,unit, in order to have

an overall weighing factor (Wi,total = Wi,meas · Wi,unit). The example given in Table III.4 clarifies

this approach. The way in which this weighing procedure influences the eventual shape of


108

the fitted distribution is explained further on in this text. Another problem at the level of

reporting is that some of the published concentrations were expressed per gram fat, without

concomitant reporting of the corresponding total fat concentration in the analyzed samples.

Such data preclude the calculation of concentrations per gram fresh weight, and were not

included in the database. Finally, a considerable amount of governmental publications

reported concentration data without mentioning the origin of the seafood (because in most of

the cases it concerns commercial samples with undefined origin), creating a lack of essential

information. But as in the end no distinction according to origin was made for most species,

these data were yet included in the database.

The fifth problem is the lack of data. It is generally known that – for many contaminants –

analytical data are often not published in open literature and therefore are difficult to access.

Some of them are available via scientific publications or via websites of national food (safety)

agencies, but in most cases these data are published in ‘grey’ reports, which are difficult to

trace and retrieve. Consequently, contaminant data could often not be found for

commercially relevant regions; e.g. no dl PCB concentration data were found for halibut

from the Northwest Atlantic Sea. Overall, for most of the seafood species, not enough data

were available to make a difference between the different fishing grounds of origin, obliging

us to group the data over the different fishing grounds. In addition, data of different species

had to be aggregated (for dioxin-like compounds and PCBs according to fat content) in cases

where no sufficient data on species level were available for the distribution fitting procedure

and similar probability distributions were determined for these grouped species (Annex

III.6-11 of this chapter). Judd et al (2003a) encountered a similar problem, when carrying out

an intake assessment of PCBs via seafood for two specific American population groups with

high seafood diets, the Squaxin Tribe and the Asian Pacific Islanders. For these population

groups, appropriate total PCB data were available for respectively 58% and 4% of the

seafood species consumed, as for such data aggregation was the only solution.

Finally, for several ecological or monitoring studies, it was not clear whether the samples

analyzed are representative for human consumption. Therefore, more detailed information

about the seafood samples is essential in order to avoid that data collected for one purpose

are incorrectly applied in other situations. Judd et al (2003b) indicated that use of data

collected in a context of ecological evaluation for the purpose of human health risk

assessments may lead to substantial risk assessment errors.


109

4.3.2. Observed variability

The result of the data collection is summarized in box plots, visualizing the observed

variability within the species (Fig. III.7-11). Species for which only a low number of data

points was available, were not plotted. Contaminant concentrations ranged per gram fresh

weight from 2.4 to 4390.0 ng for mercury, from 0.1 to 5736.6 ng for the sum of the iPCBs,

from 0.002 to 115.000 pg TEQ for the sum of dl PCBs, from 0.002 to 34.400 pg TEQ for

PCDD/Fs, and from 0.006 to 126.000 pg TEQ for the total sum of dioxin-like compounds.

4.3.2.1. Mercury

Fig. III.7 shows the published mercury data for different seafood species, sorted according to

their median mercury concentration. To protect public health, maximum levels of mercury in

fishery products are laid down by the European Commission (Commission Regulation (EC)

No 1881/2006 of 19 December 2006) (European Commission, 2006). The levels should be as

low as reasonably achievable, taking into account that for physiological reasons certain

species concentrate mercury more easily in their tissues than others. Mercury limit for fishery

products in general is 500 ng/g fresh weight. It is 1000 ng/g fresh weight for anglerfish,

Atlantic catfish, bass, blue ling, bonito, eel, halibut, little tuna, marlin, pike, plain bonito,

Portuguese dogfish, rays, redfish, sail fish, scabbard fish, shark, snake mackerel, sturgeon,

swordfish, and tuna. In the compiled database, these limits were exceeded for five species:

cod, common shrimp, tuna, swordfish, and Nile perch; for respectively 1.0%, 2.4%, 28.7%,

51.1%, and 7.7% of the concentration data.

An origin-based distinction could be made to separate the seafood coming from the

Mediterranean Sea and the other fishing grounds (possible for mackerel, squid, and sardine).

To describe the mercury content in salmon, a distinction was made between farmed salmon

and wild Pacific salmon. The box plots show that tuna and swordfish have the highest

mercury load and that mackerel, squid, and sardines from the Mediterranean Sea contain

more mercury compared to other catching areas. It was observed that mollusks and

crustacean have rather low mercury concentrations. Moreover, the box plots show the

within-species variability of the mercury concentrations, e.g. the lowest measured mercury

concentration in tuna (Fig. III.7, n°35) falls below the 75th percentile of the mercury

concentrations in halibut (species with the lowest median; Fig. III.7, n°1).


110

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36

Species

1

10

100

1000

Mer

cury

con

cent

ratio

n (n

g/g

wet

wei

ght)

- Log

sca

le

1. Halibut; 2. Squid, non-Mediterranean; 3. Tilapia; 4. Sardine, Eastern Central Atlantic Sea; 5. Mussel; 6. Scallops; 7. Oyster; 8. Norway lobster; 9. Scampi; 10. Herring; 11. Sprat; 12. Brill; 13. Salmon, farmed; 14. Mackerel, Northeast Atlantic Sea; 15. European plaice; 16. Salmon, Pacific, wild; 17. Anchovy; 18. Common shrimp; 19. Sole; 20. Crab; 21. Saithe and Pollack; 22. Haddock; 23. Redfish; 24. Trout; 25. Squid, Mediterranean Sea; 26. Cod; 27. Nile perch; 28. Skate; 29. Whiting; 30. Anglerfish; 31. Sardine, Mediterranean Sea; 32. Eel; 33. Ling; 34. Mackerel, Mediterranean Sea; 35. Tuna; 36. Swordfish

Fig. III.7 Mercury concentrations in different species, box plots. Data available from the open literature and from national and international reports

For the goal of this PhD-thesis, mercury concentrations were converted to methyl mercury,

since this organic form is the major chemical form in which mercury is present in seafood and

it is the most toxic form of the element (Clarkson & Magos, 2006). The majority of the

mercury released in the marine environment is inorganic mercury, but this can then be

converted to methyl mercury by anaerobic bacteria in sediments (Storelli et al, 2002). On the

basis of literature data, it was assumed that 80% of the total mercury in fish is present in

methylated form (Baeyens et al, 2003; Capelli et al, 2004; Easton et al, 2002; Foran et al, 2004;

Forsyth et al, 2004; Storelli et al, 2002; Storelli et al, 2003a; Storelli et al, 2003c; Storelli et al,

2005), being 68% and 34% for molluscs and crustacean (Baeyens et al, 2003; Foran et al, 2004),

respectively.

4.3.2.2. iPCBs

Fig. III.8 shows that lean seafood species have generally a lower iPCB concentration, only

crab (Fig. III.8, n°27) seems to be an exception. Large within-species variability can be

observed. The only species for which a distinction according to different catching areas was

made is herring form the Baltic Sea (Fig. III.8, n°25) versus non-Baltic (Fig. III.8, n°15), with

the first having a higher contamination load.


111

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28Species

0.1

1

10

100

1000

23457

23457

23457

23457

234579

Indi

cato

r PC

Bs (n

g/g

wet

wei

ght)

- Log

sca

le

1.Skate; 2.Scampi; 3.Scallop; 4.(Alaska) pollack and saithe; 5.Whiting; 6.Haddock; 7.FG1; 8.Cod; 9.Common shrimp; 10.European plaice; 11.FG2; 12.Sole; 13.Redfish; 14.Mussels; 15.non-Baltic Herring; 16.Halibut; 17.Sprat; 18.Squid; 19.Salmon; 20.Mackerel; 21.FG3; 22.FG4; 23.Anchovy; 24.Sardine; 25.Herring, Baltic Sea; 26.FG5; 27.Crab; 28.Eel

Fig. III.8 Concentrations of the sum of iPCBs in different species, box plots. Data available from the open literature and from national and international reports. For an explanation of FG1 to FG5 see Table III.1

4.3.2.3. dl PCBs, PCDD/Fs, total dioxin-like compounds

The European Commission published a regulation (Commission Regulation (EC) No

1881/2006 of 19 December 2006) setting maximum levels for certain contaminants in foods as

regards PCDD/Fs and dl PCBs (European Commission, 2006). Concerning muscle meat of

fish, fishery products and products thereof, the maximum levels for PCDD/Fs is 4.0 pg

TEQ/g fresh weight and the maximum level for the sum of PCDD/Fs and dl PCBs is 8.0 pg

TEQ/g fresh weight, with exception of eel which may contain 12.0 pg TEQ/g fresh weight.

In the compiled database, concentrations exceeding the limit were found for seven species.

Table III.5 gives the percentages of those data in relation to the total data.

Table III.5: Percentage of data and number of data points exceeding the limits for dioxin-like compounds Species PCDD/Fs Total dioxin-like compounds

% of the data exceeding the limit

(number of data points) Eel 4.9 (4) 24.5 (26) Halibut 2.6 (1) 10.0 (1) Herring, Baltic Sea 48.7 (127) 50.0 (93) Herring, non-Baltic 4.5 (3) 13.4 (9) Mackerel 1.6 (1) 1.9 (1) Salmon, Baltic Sea 66.7 (34) 96.0 (48) Salmon, farmed 15.4 (8) 4.5 (8) Salmon, Pacific, wild 0.0 0.0 Trout 3.3 (2) 6.3 (5) Tuna 2.4 (1) 22.0 (9)


112

Fig. III.9 visualizes that the fatty fish species have the highest dl PCB load: herring, salmon,

eel and sprat, but again with high within-species variability.

1 2 3 4 5 6 7 8 9 10111213141516171819202122232425262728293031Species

0.00

0.01

0.10

1.00

10.00

100.00

Dio

xin-

like

PCB

conc

entra

tion

(pg

TEQ

/g w

et w

eigh

t) - L

og s

cale

1.Scampi; 2.Haddock; 3.FG1; 4.Cod; 5.(Alaska) Pollack and saithe; 6.Whiting; 7.Mussels; 8.FG2; 9.Redfish; 10.Squid; 11.Common shrimp; 12.Swordfish; 13.European plaice; 14.Salmon, Pacific Ocean; 15.Trout; 16.FG4; 17.Turbot; 18.Halibut; 19.Tuna; 20.FG3; 21.Sardine; 22.Mackerel; 23.Farmed salmon; 24.Herring (North Sea); 25.Anchovy; 26.FG5; 27.Sprat; 28.Herring (Baltic Sea); 29.Eel; 30.Herring (others); 31.Salmon (Baltic Sea)

Fig. III.9 Concentrations of the sum of dl PCBs in different species, box plots. Data available from the open literature and from national and international reports. For an explanation of FG1 to FG5 see Table III.1

Fig. III.10 shows the PCDD/F concentrations of the different species of interest, ordered by

their median PCDD/F concentrations, leading to a clear gradient, but again with high

within-species variability.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36

Species

0.001

0.010

0.100

1.000

10.000

PC

DD

/F c

once

ntra

tion

(pg

TEQ

/g w

et w

eigh

t) - L

og s

cale

1.Cod, Iceland and Faroer; 2.Saithe and Pollack; 3.FG1; 4.Haddock; 5.Whiting; 6.Cod, North Sea; 7.Scampi; 8.Cod, other Nortwest Atlantic seas; 9.Swordfish; 10.Squid; 11.Redfish; 12.Mussels; 13.FG2; 14.European plaice; 15.Anchovy; 16.Trout; 17.FG4; 18.Mackerel; 19.Tuna; 20.FG3; 21.Sardines; 22.Turbot; 23.Wolf fish; 24.Halibut, Northeast Atlantic; 25.Salmon, non-Baltic; 26.Halibut, Northwest Atlantic; 27.Common shrimp; 28.Crab; 29.Herring, North Sea; 30.Eel; 31.Sole; 32.FG5; 33.Sprat; 34.Herring, Baltic Sea; 35.Salmon, Baltic Sea; 36.Herring, other Northwest Atlantic seas

Fig. III.10 PCDD/F concentrations in different species, box plots. Data available from the open literature and from national and international reports. For an explanation of FG1 to FG5 see Table III.1


113

Fig. III.11 shows the concentrations of the total sum of dioxin-like compounds (total TEQ) in

the different species. A clear gradient appears when ordering the different seafood species

according to their median total TEQ concentration. The species with the highest median

concentration are herring and salmon of the Baltic Sea (Fig. III.11, n°33, 34). In this context, it

is important to note that different methodologies can be applied to measure the

concentrations of total dioxin-like compounds, other than summing the measured PCDD/F

and dl PCB concentrations. In other words, the results in Fig. III.11 are not simply the sum of

those in Fig. III.9 and III.10.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34

Species

0.01

0.10

1.00

10.00

100.00

totT

EQ

con

cent

ratio

n (p

g TE

Q/g

fres

h w

eigh

t) - L

og s

cale

1.Haddock; 2.Scampi; 3.FG1; 4.Cod; 5.Saithe and Pollack; 6.Salmon, Pacific, wild; 7.Whiting; 8.Mussels; 9.FG2; 10.Sole; 11.Redfish; 12.Swordfish; 13.Squid; 14.European plaice; 15.Lobster; 16.Trout; 17.FG4; 18.Turbot; 19.Mackerel; 20.Scallop; 21.Common shrimp; 22.Halibut; 23.Oysters; 24.FG3; 25.Tuna; 26.Sardines; 27.Salmon, farmed; 28.Anchovy; 29.Herring, non-Baltic; 30.FG5; 31.Sprat; 32.Eel; 33.Herring, Baltic Sea; 34.Salmon, Baltic Sea

Fig. III.11 Total TEQ concentrations in different species, box plots. Data available from the open literature and from national and international reports. For an explanation of FG1 to FG5 see Table III.1

For the dl PCBs and the total TEQ concentration, a distinction according to fishing ground

could only be made for the species herring and salmon, making clear that the Baltic herring

and salmon are more contaminated (Fig. III.9 and III.11). A distinction between different

fishing grounds could be made for the PCDD/F concentration distributions for cod, halibut,

herring, and salmon (Fig. III.10).


114

4.3.3. Distribution fitting

For each contaminant, the concentration data together with their cumulative probabilities of

occurrence, F(xi), were used per species and – where possible - per fishing ground, in order

to characterize distributions using BestFit® Version 4.5 (see Table III.4 given previously in

combination with Fig. III.12 and explanatory notes later on). On the basis of the suggested

distributions and goodness-of-fit test statistics delivered by BestFit®, it has been decided

which kind of distribution with matching parameters could be applied to adequately

describe the data. Annex III.6 to III.11 give the selected distributions for all different species

and the five contaminants.

Annex III.6 and III.7 give the distributions and parameters of the mercury and methyl

mercury concentration in different fish species. The tables show that a distinction according

to origin of the seafood product and its mercury concentration could be made for mackerel,

sardine, squid, and salmon. The percentages given in the second column are a result of the

work executed in the traceability study and give an indication of the importance of each

catching area (part 2 of this chapter). The data in the table also indicate that no data were

available for three seafood products: caviar, conger, and surimi. In those cases, the mercury

contamination was assumed to be negligible. The underestimation of the total mercury

intake caused by this assumption will be very small, due to the low amount of these

products that is consumed.

Annex III.8 gives the distributions and parameters of the iPCB concentration in different fish

species. As a solution to fill the gaps of lacking data, the different concentrations data of all

species belonging to the same fat group were pooled and used for distribution fitting. When

no data were available for an individual species, the distribution of the corresponding fat

group was used. This approach was also applied for dl PCBs, PCDD/Fs, and total dioxin-

like compounds.

Annex III.9 to III.11 present the distributions and parameters of the different fish species for

the different dioxin-like compounds. Also here, aggregation per fat group was used to fill

data gaps.


115

Overall, six different distribution models were selected: a normal distribution, a lognormal

distribution, a logistic distribution, a loglogistic distribution, a 4-parametrical beta

distribution, and a uniform distribution. Since the normal, lognormal, logistic, and loglogistic

distributions have open ends to infinity, they were in this application truncated to avoid

unrealistic high or low concentrations. However, it is difficult to quantify biologically

plausible lower and upper bounds. Yet, they can be higher than the highest value in the

available database, especially when the quantity of data is rather limited. In general,

maximum values in the data are not suitable upper bounds, and it seems inappropriate to

exclude the possibility of more extreme values (Paulo et al, 2006). As was done for the

nutrient distributions, the imposed minimum was defined as half of the lowest observed

concentration and the imposed maximum as double of the highest observed concentration. A

uniform distribution was applied in the cases where only one or two concentration data were

available and no possibility for aggregation existed (in the case of two available data points,

they were used as minimum and maximum of the uniform distribution; in the case of only

one available data point, the minimum and maximum were defined as respectively the

concentration minus/plus half of that concentration). Due to the uncertainty related to the

low number of data points, the latter approach was only applied when no alternative was

available.

Fig. III.12 to III.14 show some examples of the selected distributions to describe the

contaminant concentration for a certain seafood species from a certain fishing ground, as

well as the original data points. Both representations together give a visual idea of the

goodness of fit.

Fig. III.12 (together with Table III.4 given previously) indicated that more data points were

available for the mercury concentration in sardines from the Eastern Central Atlantic Sea

than from the Mediterranean Sea. Moreover, it is clear that the latter has the highest mercury

contamination. The sudden rise in the Mediterranean Sea data points (from 75.5 ng/g with a

cumulative probability of 0.10 to 130.0 ng/g with a cumulative probability of 0.94) is caused

by the high weight of the latter caused by the high number of measurements (Wi,meas) behind

that concentration (Table III.4).


116

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

0 25 50 75 100 125 150 175 200 225 250Mercury concentration in ng/g wet weight

Cum

ulat

ive

prob

abili

ty

Eastern Central Atlantic Sea (fitted)

Mediterranean Sea (fitted)

Eastern Central Atlantic Sea (data points)

Mediterranean Sea (data points)

Fig. III.12 Used data points and cumulative probability function of the mercury concentration in sardines from the Eastern Central Atlantic Sea and the Mediterranean Sea; detailed information about the data points is given

in Table III.4

Fig. III.13 shows a case where a rather low number of data points was available.

Consequently, each data point has a high influence on the final distribution and parameter

selection.

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

0 20 40 60 80 100 120

iPCB concentration in pg/g wet weight

Cum

ulat

ive

prob

abili

ty

Data pointsFitted

Fig. III.13 Used data points and cumulative probability function of the iPCB concentration in anchovy (no

distinction was made between different origins)


117

Fig. III.14 shows the PCDD/F concentration in herring from two different fishing grounds,

showing a higher contamination level in Baltic herring as compared to the North Sea.

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

0 5 10 15 20 25 30

PCDD/F concentration (pg TEQ/g wet weight)

Cum

ulat

ive

prob

abili

ty

Baltic Sea (fitted)Baltic Sea (data points)North Sea (fitted)North Sea (data points)

Fig. III.14 Used data points and cumulative probability functions and of the PCDD/F concentration in herring

from the Baltic Sea and the North Sea

To assess the effect of the use of the weighing factor, all the distributions were fitted a second

time, but without weighing the concentration data. In this second approach, every data point

was given the same probability of occurrence, regardless of the number of sample units per

sample and the total number of samples behind the considered measured concentration. It is

hard to decide which results are better, as no “golden standard” exists. But in general,

distributions predicting higher concentrations were determined in the non-weighing

approach. This is illustrated by Fig. III.15 and III.16. The most probable explanation is that

most of the high concentrations measured are not mean values, but rather the result of an

individual extreme measurement or a measurement based on a small number of samples. In

the weighing approach, a small probability of occurrence will be attributed to these “hot

spot” concentrations. In contrast, in the non-weighing approach, these high concentrations

will have the same probability of occurrence than mean concentrations based on a large

number of samples, leading to distributions predicting higher concentrations. On the other

hand, the same will count for extremely low observations, but the extreme lows are not as

extreme as the extreme highs. As a result, using distributions resulting from the non-

weighing approach in a probabilistic intake assessment will lead to higher assessed intakes,

closer to toxicological thresholds.


118

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

0 20 40 60 80 100 120 140 160 180

Mercury concentration in ng/g wet weight

Cum

ulat

ive

prob

abili

ty

With weighing factor

Without weighing factor

Fig. III.15 Determined distributions for the mercury concentration in trout, with and without using the

weighing factor

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

0 5 10 15 20 25 30 35PCDD/F concentration in pg TEQ/g wet weight

Cum

ulat

ive

prob

abili

ty

With weighing factor

Without weighing factor

Fig. III.16 Determined distributions for the PCDD/F concentration in trout, with and without use of the

weighing factor

4.4. Conclusion

An extended contamination database of seafood was established, covering relevant seafood

species from several relevant fishing grounds for human consumption in Belgium. However,

several problems were encountered during the establishment of the database, related to a

lack of international recommendations to standardize analytical methodologies and the

format for reporting the results. Harmonization of sampling plans, sample handling,

analytical methodologies, and the format of results reporting would improve the usefulness

of published contamination data. Our conclusions confirm those earlier made by the EU


119

Scientific Steering Committee about the need to improve the comparability of data critical to

conduct human intake assessments (EU Scientific Steering Committee, 2000). Because of

current lack of established guidance regarding methods dealing with uncertainty in intake

assessment and probabilistic approaches, the EFSA will consider further work in this area, by

contributing to the development of a European framework for the harmonization of food-

related data collection in the EU and making these data publicly accessible (EFSA, 2005a). In

order to solve the current encountered problems, a weighing method has been drawn up,

applied, and compared in order to determine adequately contamination distributions as

close to reality as possible. As such, a probabilistic assessment regarding the intake of

contaminants via seafood consumption can be executed on the basis of these distributions.


120

5. Annexes of chapter III ANNEX III.1 - Nomenclatural table of 41 seafood species, relevant for Belgian consumption English name Dutch name French name Scientific name Anchovy Ansjovis Anchois Engraulis encrasicolus Anglerfish Zeeduivel, lotte Badroie, lotte,

Crapaud Lophius piscatorius

Brill Griet Barbue Scophthalmus rhombus Cod Kabeljauw Cabillaud Gadus morhua Common (brown) shrimp

Noordzeegarnaal Crevette grise Crangon crangon

Common whelk Slak/Wulk Buccin Buccinidae Conger Zeepaling,

congeraal Congre Conger conger

Crab Krab Crabe Cancer pagurus Eel Paling Anguille Anguilla anguilla European catfish Meerval Silure, poisson-chat Clarias gariepinus European plaice Schol, pladijs Plie Pleuronectes platessa Haddock Schelvis Eglefin Melanogrammus aeglefinus Halibut Heilbot Flétan Hippoglossus hippoglossus/stenolepis –

Reinhardtius hippoglossoides Herring Haring Hareng Clupea harengus John dory Zonnevis Saint-pierre Zeus faber Ling Leng Lingue Molva molva/dypterygia Lobster Zeekreeft Homard Homarus gammarus Mackerel Markeel Maquereau Scomber scombrus Milkfish Melkvis, bandeng Chanos Chanos chanos Mussel Mossel Moule Mytilus edulis Nile perch Victoriabaars Perche du Nil Lates niloticus Norway lobster Langoestine Langoustine Nephrops norvegicus Oyster Oesters Huître Ostrea edulis - Crassostrea gigas Redfish Roodbaars Sébaste Sebastes marinus/mentella Saithe &Pollack Alaska koolvis Lieu de l'Alaska Theregra chalcogramma Saithe &Pollack Koolvis & Pollack Lieu noir/jeune Pollachius pollachius/virens Salmon Zalm, Atlantische Saumon Salmo salar Salmon Zalm, Pacifische Saumon Oncorhynchus spp Sardine, pilchard Sardien Sardine, pilchard Sardina pilchardus Scampi Scampi,

tijgergarnaal, gamba Crevette géante, tigrée Penaeus spp

Sea bream Zeebrasem, dorade Dorade Pagellus bogaraveo Skate, ray Rog Raie Rajidae spp. Sole (Dover) Tong Sole (commune) Solea solea Sprat Sprot Sprat, amelette Sprattus sprattus Squid, octopus Inktvis, octopus Poulpe, encornet Octopus vulgaris Squid, octopus Inktvis, pijlinktvis Calmar Loligo forbesi/vulgaris Squid, octopus Inktvis, zeekat Sèche Sepia officinalis Scallop Sint-Jakobsschelp Coquille Saint-Jacques Pecten maximus/jacobeus Swordfish Zwaardvis Espadon Xiphias gladius Tilapia Tilapia Tilapia Oreochromis niloticus/aureaus/mossambica Trout Forel Truite Salmo trutta Trout, rainbow Forel, regenboog- Truite arc-en-ciele Oncorhynchus mykiss Tuna Tonijn Thon Thunnus albacares/alalunga/maccoyii/

obesus/thynnus – Katsuwonus pelamis

Turbot Tarbot Turbot Scophthalmus maximus, Psetta maxima Whiting Wijting Merlan Merlangius merlangus Wolf fish Zeewolf Loup de mer Anarhichas lupus


121

ANNEX III.2 - Distribution and its parameters(Param) for the EPA&DHA concentration in relevant seafood species, as well as the number of data points used (N) Species Distribution Param1 Param2 Param3 Param4 Param5 N uniform min max betageneral α1 α2 min max normal µ σ Tr min Tr max loglogistic β α Tr min Tr max γ (min) lognormal µ σ Tr min Tr max shift Anchovy uniform NA NA 3.0852 20.6615 NA 7 Anglerfish normal 1.5832 0.8678 0.5000 5.2200 NA 5 Caviar normal 15.2756 13.1798 0.0000 NA NA 9 Cod normal 3.3168 0.9774 0.5000 8.8000 NA 29 Common shrimp normal 3.6600 0.9785 1.5000 14.0000 NA 3 Common whelk betageneral 0.3113 2.9108 0.0692 1.0713 NA 3 Conger uniform NA NA 0.0000 4.0000 NA 2 Crab betageneral 0.9539 3.6449 2.3222 9.5440 NA 21 Eel betageneral 0.5414 2.3157 0.0000 40.3848 NA 10 European plaice normal 2.5188 2.5569 0.5000 10.0200 NA 6 Haddock betageneral 0.5221 2.7852 1.6850 2.9060 NA 9 Halibut normal 4.7376 3.6458 0.2450 23.5600 NA 14 Herring loglogistic 16.1697 10.6812 3.4350 56.0000 -3.5261 37 John dory uniform NA NA 2.5000 7.5000 NA 1 Lobster normal 1.7130 1.6039 0.0000 NA NA 6 Mackerel betageneral 0.3500 2.3370 12.0212 45.1811 NA 18 Mussel loglogistic 56.5106 42.5258 0.9840 16.2200 -53.1307 11 Nile perch betageneral 0.2856 2.7272 0.2626 10.5809 NA 11 Norway lobster normal 1.7130 1.6039 0.0000 NA NA 6 Saithe and Pollack betageneral 1.0133 2.4937 1.7505 10.4212 NA 12 Salmon normal 20.6855 8.0150 2.2500 65.4000 NA 67 Sardine normal 19.9393 10.9294 1.9730 NA NA 45 Scallop loglogistic 0.0484 1.0334 0.9450 7.4000 1.8880 6 Scampi normal 4.1746 0.8729 1.0000 10.9000 NA 10 Sea bream normal 4.4740 3.9125 0.0000 17.7000 NA 8 Skate uniform NA NA 1.0000 2.5500 NA 2 Sole betageneral 0.2856 2.7272 0.2626 10.5809 NA 11 Sprat loglogistic 22.6204 2.9426 6.5000 71.4600 -9.4869 3 Squid normal 3.5102 1.7501 0.5250 18.2000 NA 9 Surimi normal 3.5900 0.2735 1.6000 NA NA 3 Trout lognormal 13.9674 2.7366 1.0000 48.2400 -4.2529 33 Tuna betageneral 0.3667 2.0300 2.1431 34.7867 NA 20 Whiting normal 2.3234 0.5577 0.5000 10.3600 NA 4 Wolffish normal 4.4740 3.9125 0.0000 17.7000 NA 8


122

ANNEX III.3 - Distribution and its parameters (Param) for the vitamin D concentration in relevant seafood species, as well as the number of data points used (N) Species Distribution Param1 Param2 Param3 Param4 Param5 N uniform min max betageneral α1 α2 min max normal µ σ Tr min Tr max lognormal µ σ Tr min Tr max shift Anchovy lognormal 0.0486 0.0757 0.0000 0.2800 -0.0039 12 Anglerfish uniform NA NA 0.0000 0.0010 NA 2 Caviar betageneral 0.6357 2.2928 0.0191 0.3793 NA 16 Cod lognormal 0.0300 0.0817 0.0000 0.1560 -0.0022 16 Common shrimp uniform NA NA 0.0000 0.0010 NA 3 Common whelk uniform NA NA 0.0000 0.0010 NA 1 Conger uniform NA NA 0.0000 0.0010 NA 3 Crab uniform NA NA 0.0000 0.0010 NA 2 Eel lognormal 0.1574 0.3891 NA 1.4000 0.0196 16 European plaice lognormal 0.0013 0.0121 0.0000 0.0600 0.0000 6 Haddock lognormal 0.0015 0.0015 0.0000 0.0280 -0.0004 6 Halibut betageneral 0.3895 0.6429 0.0088 0.1526 NA 11 Herring normal 0.1477 0.0909 0.0000 0.8140 NA 60 John dory uniform NA NA 0.0000 0.0010 NA 1 Lobster uniform NA NA 0.0000 0.0010 NA 2 Mackerel lognormal 0.1127 0.0329 0.0005 0.4220 -0.0608 44 Mussel lognormal 0.0024 0.0084 0.0000 0.0276 -0.0001 8 Nile perch uniform NA NA 0.0000 0.0010 NA 1 Norway lobster uniform NA NA 0.0000 0.0010 NA 1 Saithe and Pollack normal 0.0091 0.0090 0.0000 0.0500 NA 9 Salmon lognormal 0.1383 0.0669 0.0000 0.6000 NA 58 Sardine lognormal 0.2484 0.0629 0.0050 0.6000 -0.1633 34 Scallop uniform NA NA 0.0000 0.0010 NA 1 Scampi normal 0.0006 0.0003 NA 0.0100 NA 4 Sea bream normal 0.0070 0.0061 0.0000 0.0460 NA 4 Skate uniform NA NA 0.0000 0.0010 NA 1 Sole lognormal 0.0522 0.0442 0.0000 0.1813 -0.0352 5 Sprat normal 0.1477 0.0909 0.0000 0.8140 NA 3 Squid uniform NA NA 0.0000 0.0010 NA 5 Surimi uniform NA NA 0.0000 0.0010 NA 0 Trout lognormal 3.4966 0.0267 0.0000 0.3800 -3.4261 28 Tuna lognormal 0.0343 0.0240 0.0000 0.4600 -0.0039 31 Whiting uniform NA NA 0.0000 0.0010 NA 11 Wolffish betageneral 0.2263 0.4041 0.0049 0.0160 NA 4


123

ANNEX III.4 - Distribution and its parameters (Param) for the iodine concentration in relevant seafood species, as well as the number of data points used (N) Species Distribution Param1 Param2 Param3 Param4 Param5 N uniform min max betageneral α1 α2 min max normal µ σ Tr min Tr max logistic α β Tr min Tr max loglogistic β α Tr min Tr max γ (min) Anchovy betageneral 0.4037 1.9862 0.0036 1.8848 NA 15 Anglerfish loglogistic 5.3796 5.5783 0.002 21.26 -3.1012 44 Caviar normal 0.4617 0.3209 0.002 2.798 NA 3 Cod logistic 2.3563 0.3901 0.145 21.26 NA 19 Common shrimp loglogistic 0.1989 1.1331 NA 14 0.1551 29 Common whelk loglogistic 0.1989 1.1331 NA 14 0.1551 29 Conger betageneral 0.4037 1.9862 0.0036 1.8848 NA 15 Crab loglogistic 5.3796 5.5783 0.002 21.26 -3.1012 44 Eel loglogistic 0.1138 1.9765 0.005 1.6 -0.0199 5 European plaice logistic 0.3322 0.0044 0.05 3.8 NA 10 Haddock logistic 1.449 0.5608 0.12 5.34 NA 12 Halibut betageneral 0.2566 0.963 0.001 1.2251 NA 5 Herring logistic 0.3361 0.0528 0.1215 1.33 NA 13 John dory loglogistic 0.1989 1.1331 NA 14 0.1551 29 Lobster normal 1.19 0.75 0.5 14 NA 3 Mackerel logistic 0.7664 0.0362 0.22 3.02 NA 11 Mussel normal 1.3081 0.0317 0.525 2.8 NA 3 Nile perch uniform NA NA 0.002 0.006 NA 2 Norway lobster loglogistic 0.1989 1.1331 NA 14 0.1551 29 Saithe and Pollack loglogistic 1.1553 3.5795 0.002 4.7 -0.4792 10 Salmon normal 0.3106 0.0637 0.15 1.3 NA 5 Sardine uniform NA NA 0.145 0.48 NA 3 Scallop loglogistic 5.3796 5.5783 0.002 21.26 -3.1012 44 Scampi loglogistic 0.1989 1.1331 NA 14 0.1551 29 Sea bream betageneral 0.4037 1.9862 0.0036 1.8848 NA 15 Skate loglogistic 5.3796 5.5783 0.002 21.26 -3.1012 44 Sole normal 0.1773 0.0229 0.002 0.5 NA 3 Sprat betageneral 0.7072 2.4604 0.2206 1.6409 NA 23 Squid loglogistic 0.1989 1.1331 NA 14 0.1551 29 Surimi loglogistic 5.3796 5.5783 0.002 21.26 -3.1012 44 Trout normal 0.1309 0.0668 0.0752 0.5 NA 8 Tuna normal -0.0882 0.4142 0.035 1.12 NA 6 Whiting loglogistic 5.3796 5.5783 0.002 21.26 -3.1012 44 Wolffish betageneral 0.4037 1.9862 0.0036 1.8848 NA 15


124

ANNEX III.5 - Distribution and parameters (Param) for the fat concentration in relevant seafood species, as well as the number of data points used (N) Species Distribution Param1 Param2 Param3 Param4 Param5 N uniform min max normal µ σ Tr min Tr max lognormal µ σ Tr min Tr max shift logistic α β Tr min Tr max Anchovy logistic 35.7088 10.1068 11.5000 160.0000 NA 10 Anglerfish lognormal 4.3921 5.1988 2.0000 39.0000 3.4499 15 Caviar lognormal 80.4766 73.9635 9.5000 424.0000 8.5785 32 Cod lognormal 19.2092 2.3810 0.0000 34.0000 -12.3241 60 Common shrimp uniform NA NA 4.1370 24.5890 NA 9 Common whelk normal 8.6198 2.7826 2.0000 24.0000 NA 4 Conger lognormal 99.0803 32.8900 3.5000 228.0000 -46.9713 8 Crab lognormal 28.9413 4.9991 2.5000 110.0000 -20.5289 40 Eel normal 213.6286 68.2837 35.5000 660.0000 NA 28 European plaice uniform NA NA 4.5910 20.4270 NA 31 Haddock lognormal 3.3519 2.1557 2.3500 42.0000 3.8266 19 Halibut logistic 33.1170 16.2834 0.5000 354.8000 NA 28 Herring logistic 123.7569 20.0719 6.0000 662.0000 NA 82 John dory normal 7.6881 5.1868 1.5000 28.0000 NA 4 Lobster normal 12.9476 3.7172 2.9500 38.0000 NA 21 Mackerel lognormal 272.8077 65.2393 55.0000 618.0000 -89.6701 58 Mussel lognormal 73.5831 10.1289 3.0000 89.6000 -51.2564 21 Nile perch normal 13.7534 6.8617 4.0000 48.0000 NA 4 Norway lobster uniform NA NA 3.5000 16.0000 NA 3 Saithe and Pollack lognormal 22.6623 3.0536 1.5000 40.0000 -13.1067 30 Salmon normal 98.1080 35.8896 16.0000 361.2000 NA 100 Sardine logistic 137.6906 25.5402 24.0000 368.0000 NA 56 Scallop lognormal 12.8756 5.3110 0.5000 28.0000 -5.7020 7 Scampi normal 12.9792 5.4646 3.0000 50.0000 NA 24 Sea bream uniform NA NA 14.0000 70.0000 NA 4 Skate normal 7.0500 3.6473 1.0000 25.2000 NA 12 Sole lognormal 40.7345 6.0389 0.6500 52.0000 -35.0947 22 Sprat normal 130.1250 41.4813 2.9000 36.8000 NA 8 Squid lognormal 14.5848 12.6205 2.0000 94.0000 2.2428 29 Surimi logistic 8.1651 0.7409 2.0000 26.2000 NA 8 Trout lognormal 57.7602 44.4512 50.2500 366.0000 8.7146 56 Tuna lognormal 93.5312 34.8306 1.0000 310.0000 -54.9499 34 Whiting logistic 5.8661 0.5255 1.5000 16.2000 NA 18 Wolffish logistic 31.1497 6.6996 2.9500 118.0000 NA 19


125

ANNEX III.6 - Distribution and parameters (Param) for the mercury concentration in relevant seafood species, as well as the number of data points used (N) Species % Distribution Param1 Param2 Param3 Param4 Param5 N uniform min max betageneral α1 α2 min max normal µ σ Tr min Tr max lognormal µ σ Tr min Tr max shift logistic α β Tr min Tr max loglogistic β α Tr min Tr max γ (min) Anchovy 100 normal 53.2489 18.1537 14.0000 208.0000 NA 13 Anglerfish 100 uniform NA NA 21.3641 1141.271 NA 14 Caviar 100 no_distribution NA NA NA NA NA 0 Cod 100 logistic 66.6564 6.7232 7.0000 8780.000 NA 104 Common shrimp 100 loglogistic 34.6900 3.0850 10.0000 1134.000 2.8172 42 Common whelk 100 uniform NA NA 50.5000 151.5000 NA 1 Conger 100 no_distribution NA NA NA NA NA 0 Crab 100 loglogistic 65.1614 7.0722 8.5000 300.0000 -1.9272 11 Eel 100 normal 95.8212 71.4331 5.0000 640.0000 NA 27 European plaice 100 lognormal 147.393 22.9030 7.5000 404.0000 -100.71 41 Haddock 100 betageneral 1.2056 2.9272 8.1556 151.1400 NA 11 Halibut 100 normal 82.1316 7.8708 34.5000 308.0000 NA 3 Herring 100 logistic 33.2401 12.7331 1.5000 230.0000 NA 15 John dory 100 uniform NA NA 20.0000 75.0000 NA 2 Lobster 100 betageneral 0.3575 0.3728 59.7916 454.5314 NA 7 Mackerel, Mediterranean Sea

1.8 lognormal 262.450 32.3860 63.0000 580.0000 -57.2340 5

Mackerel, Northeast Atlantic Ocean

98.2 logistic 29.8991 8.6537 1.5000 200.0000 NA 12

Mussel 100 betageneral 0.0906 4.3860 8.4665 156.4585 NA 13 Nile perch 100 normal 82.3325 41.0294 4.8500 1181.600 NA 13 Norway lobster 100 betageneral 0.3575 0.3728 59.7916 454.5314 NA 7 Saithe and Pollack 100 logistic 47.6786 7.5200 3.5000 500.0000 NA 15 Salmon, farmed 64.07 betageneral 0.7041 2.8444 16.9038 81.5676 NA 42 Salmon, Pacific Ocean

35.93 logistic 49.4647 13.8422 5.0000 234.0000 NA 32

Sardine, Eastern Central Atlantic Ocean

60.62 lognormal 15.8642 9.5737 NA 208.0000 5.7499 18

Sardine, Mediterranean Sea

39.38 logistic 101.107 11.5022 23.3650 468.0000 NA 8

Scallop 100 loglogistic 0.3039 1.0202 NA 84.0000 19.9561 9 Scampi 100 betageneral 0.0624 2.9870 17.0001 266.2788 NA 25 Sea bream 100 uniform NA NA 25.0000 75.0000 NA 1 Skate 100 uniform NA NA 4.6500 1650.000 NA 24 Sole 100 betageneral 0.7578 2.3677 4.4894 363.8900 NA 32 Sprat 100 uniform NA NA 5.0000 90.0000 NA 4 Squid, Mediterranean Sea

89.82 logistic 11.5608 1.5213 1.5000 166.8000 NA 9

Squid, Southwest Atlantic Ocean

10.18 uniform NA NA 35.0000 133.5000 NA 3

Surimi 100 no_distribution NA NA NA NA NA 0 Trout 100 lognormal 22.3350 20.2067 NA 298.0000 29.0953 27 Tuna 100 lognormal 658.790 892.639 NA 6460.000 15.6585 153 Whiting 100 lognormal 54.2950 46.8510 NA 588.0000 22.5420 39 Wolffish 100 uniform NA NA 25.0000 75.0000 NA 1


126

ANNEX III.7 - Distribution and parameters (Param) for the calculated methyl mercury concentration in relevant seafood species, as well as the number of data points used (N) Species % Distribution Param1 Param2 Param3 Param4 Param5 N uniform min max betageneral α1 α2 min max normal µ σ Tr min Tr max lognormal µ σ Tr min Tr max shift logistic α β Tr min Tr max loglogistic β α Tr min Tr max γ (min) Anchovy 100 normal 42.5055 14.4910 11.1754 245.8584 NA 13 Anglerfish 100 betageneral 0.3877 0.1551 63.5953 606.7781 NA 14 Caviar 100 no_distribution NA NA NA NA NA 0 Cod 100 logistic 53.2079 5.3668 5.5877 7008.5602 NA 104 Common shrimp 100 loglogistic 11.7632 3.0850 NA 384.5291 0.9553 42 Common whelk 100 uniform NA NA 34.3575 103.0725 NA 1 Conger 100 no_distribution NA NA NA NA NA 0 Crab 100 loglogistic 22.0956 7.0722 2.8823 101.7273 -0.6535 11 Eel 100 normal 76.4884 57.0208 3.9912 510.8745 NA 27 European plaice 100 loglogistic 15.2829 1.6400 NA 322.4896 17.8938 41 Haddock 100 betageneral 1.2056 2.9272 6.5101 120.6478 NA 11 Halibut 100 normal 65.5609 6.2828 27.5393 245.8584 NA 3 Herring 100 logistic 26.5336 10.1641 1.1974 183.5955 NA 15 John dory 100 uniform NA NA 15.9648 59.8681 NA 2 Lobster 100 betageneral 0.3575 0.3728 20.2748 154.1275 NA 7 Mackerel, Mediterranean Sea

1.8 lognormal 209.5021 25.8515 50.2892 462.9801 -45.6865 5

Mackerel, Northeast Atlantic Ocean

98.2 logistic 23.8667 6.9077 1.1974 159.6483 NA 12

Mussel 100 betageneral 0.0906 4.3860 2.8709 53.0536 NA 13 Nile perch 100 normal 65.7212 32.7514 3.8715 943.2021 NA 13 Norway lobster 100 betageneral 0.3575 0.3728 20.2748 154.1275 NA 7 Saithe and Pollack 100 logistic 38.0590 6.0028 2.7938 399.1207 NA 15 Salmon, farmed 64.1 betageneral 0.7041 2.8444 13.4934 65.1106 NA 42 Salmon, Pacific Ocean

35.9 lognormal 201.7383 18.1925 3.9912 186.7885 -161.618 32

Sardine, Eastern Central Atlantic Ocean

60.6 loglogistic 20.7045 6.4354 1.9956 208.0000 -5.1385 18

Sardine, Mediterranean Sea

39.4 logistic 80.7084 9.1815 18.6509 373.5770 NA 8

Scallop 100 loglogistic 0.1030 1.0202 NA 28.4836 6.7669 9 Scampi 100 lognormal 34.1832 40640.102 NA 168.8673 5.7644 25 Sea bream 100 uniform NA NA 19.9560 59.8681 NA 1 Skate 100 betageneral 0.5848 1.3428 17.2469 1217.7227 NA 24 Sole 100 betageneral 0.7578 2.3677 3.5836 290.4710 NA 32 Sprat 100 uniform NA NA 3.9912 71.8417 NA 4 Squid, Mediterranean Sea

10.2 uniform NA NA 23.8121 90.8263 NA 3

Squid, Southwest Atlantic Ocean

89.8 logistic 7.8653 1.0350 1.0205 113.4818 NA 9

Surimi 100 no_distribution NA NA NA NA NA 0 Trout 100 lognormal 17.8287 16.1298 NA 237.8760 23.2251 27 Tuna 100 betageneral 0.6326 3.3760 80.1298 2628.8888 NA 153 Whiting 100 lognormal 43.3407 37.3987 NA 469.3660 17.9942 39 Wolffish 100 uniform NA NA 19.9560 59.8681 NA 1


127

ANNEX III.8 - Distribution and parameters (Param) for the iPCB concentration in relevant seafood species, as well as the number of data points used (N) Species % Distribution Param1 Param2 Param3 Param4 Param5 N uniform min max betageneral α1 α2 min max normal µ σ Tr min Tr max lognormal µ σ Tr min Tr max shift logistic α β Tr min Tr max loglogistic β α Tr min Tr max γ (min) Anchovy 100 betageneral 0.5762 3.0084 13.2207 92.6398 NA 10 Anglerfish 100 uniform NA NA 0.1000 1.8000 NA 3 Caviar 100 loglogistic 11.0376 2.5872 NA 203.0000 9.6586 21 Cod 100 logistic 0.9595 0.2620 0.1000 70.0000 NA 57 Common shrimp 100 loglogistic 0.7017 0.6924 NA 392.0000 0.2917 45 Common whelk 100 lognormal 10.6747 21.5897 0.1500 392.0000 -0.5751 120 Conger 100 loglogistic 11.0376 2.5872 NA 203.0000 9.6586 21 Crab 100 loglogistic 60.5177 1.5296 0.9950 560.0000 -24.0733 13 Eel 100 lognormal 386.1751 2410.4067 NA 11472.6680 9.4147 159 European plaice 100 loglogistic 1.0512 1.8705 0.1500 70.0000 0.1442 31 Haddock 100 loglogistic 0.4932 2.2905 0.1000 4.6763 -0.1356 11 Halibut 100 betageneral 1.9509 0.4589 9.8699 22.1100 NA 6 Herring, Baltic Sea 5.44 betageneral 1.0979 3.7898 4.7632 193.0384 NA 156 Herring, others 94.56 betageneral 0.3269 5.1964 7.0347 68.5782 NA 5 John dory 100 lognormal 10.6747 21.5897 0.1500 392.0000 -0.5751 120 Lobster 100 lognormal 10.6747 21.5897 0.1500 392.0000 -0.5751 120 Mackerel 100 lognormal 8.3498 10.4416 NA 116.0000 3.6131 16 Mussel 100 normal 7.4274 6.0454 0.5800 44.6200 NA 13 Nile perch 100 lognormal 10.6747 21.5897 0.1500 392.0000 -0.5751 120 Norway lobster 100 lognormal 10.6747 21.5897 0.1500 392.0000 -0.5751 120 Saithe and Pollack 100 uniform NA NA 0.3691 1.8667 NA 3 Salmon 100 loglogistic 91.3367 21.7584 0.1523 640.7226 -79.8300 90 Sardine 100 logistic 13.6936 3.6041 2.6635 208.0000 NA 10 Scallop 100 logistic 0.8816 0.3054 0.1700 35.0000 NA 6 Scampi 100 uniform NA NA 0.0425 1.0990 NA 4 Sea bream 100 loglogistic 11.0376 2.5872 NA 203.0000 9.6586 21 Skate 100 betageneral 0.0682 0.9075 0.3674 38.0536 NA 13 Sole 100 lognormal 11.5521 12.1603 0.1900 70.0000 -4.4486 19 Sprat 100 logistic 9.1463 1.3977 3.5000 65.3874 NA 6 Squid 100 uniform NA NA 0.2900 66.2355 NA 6 Surimi 100 logistic 0.9959 0.4176 0.0500 560.0000 NA 120 Trout 100 logistic 13.3240 3.8499 2.6635 208.0000 NA 12 Tuna 100 loglogistic 11.0376 2.5872 NA 203.0000 9.6586 21 Whiting 100 lognormal 5.7108 13.9496 0.0500 70.0000 -0.1728 28 Wolffish 100 loglogistic 11.0376 2.5872 NA 203.0000 9.6586 21


128

ANNEX III.9 - Distribution and parameters (Param) for the dl PCB concentration in relevant seafood species, as well as the number of data points used (N)

Species % Distribution Param1 Param2 Param3 Param4 Param5 N uniform min max betageneral α1 α2 min max normal µ σ Tr min Tr max lognormal µ σ Tr min Tr max shift logistic α β Tr min Tr max Anchovy 100 uniform NA NA 0.25 11.85 NA 4 Anglerfish 100 loglogistic 0.0747 1.2186 0.0014 2.8 0.0002 94 Caviar 100 loglogistic 1.1167 1.598 0.01 46.1106 -0.319 68 Cod 100 lognormal 0.1891 0.3882 NA 1.064 0.0232 43 Common shrimp 100 normal 0.4227 0.3025 0.005 1.86 NA 5 Common whelk 100 lognormal 1.2752 0.3955 0.0008 4.8 -0.7468 65 Conger 100 loglogistic 1.1167 1.598 0.01 46.1106 -0.319 68 Crab 100 loglogistic 0.0747 1.2186 0.0014 2.8 0.0002 94 Eel 100 lognormal 11.4557 20.5263 0.045 230 0.0159 142 European plaice 100 logistic 0.3374 0.0465 0.063 2.078 NA 14 Haddock 100 loglogistic 0.0227 2.5595 NA 0.732 0.0177 31 Halibut 100 normal 1.1586 0.5388 0.215 23.32 NA 10 Herring, Baltic Sea 5.44 lognormal 6.3217 3.0684 0.2403 74.98 -2.1033 182 Herring, North Sea 50.89 betageneral 0.2291 2.0719 0.89 6.5646 NA 34 Herring, others 43.67 betageneral 0.4297 1.2207 4.9682 12.1607 NA 4 John dory 100 lognormal 1.2752 0.3955 0.0008 4.8 -0.7468 65 Lobster 100 normal 0.1695 0.3498 0.0021 1.08 NA 3 Mackerel 100 loglogistic 2.8421 13.5998 0.18 41.18 -1.6156 48 Mussel 100 normal 0.5565 1.2006 0.05 4.06 NA 13 Nile perch 100 lognormal 1.2752 0.3955 0.0008 4.8 -0.7468 65 Norway lobster 100 lognormal 1.2752 0.3955 0.0008 4.8 -0.7468 65 Saithe and Pollack 100 betageneral 0.7251 1.7462 0.0029 0.8 NA 4 Salmon, Baltic Sea 45.08 loglogistic 3.0346 3.3585 NA 31.4 6.1069 46 Salmon, farmed 27.62 lognormal 3.6574 1.517 0.035 16 -1.9129 162 Salmon, Pacific Ocean 27.30 betageneral 0.4565 1.7419 0.0076 4.2119 NA 14 Sardine 100 logistic 2.0999 1.3447 0.001 12.6 NA 8 Scallop 100 loglogistic 0.0747 1.2186 0.0014 2.8 0.0002 94 Scampi 100 betageneral 1.5552 7.3677 0.0011 0.2469 NA 16 Sea bream 100 loglogistic 1.1167 1.598 0.01 46.1106 -0.319 68 Skate 100 loglogistic 0.0747 1.2186 0.0014 2.8 0.0002 94 Sole 100 lognormal 1.2752 0.3955 0.0008 4.8 -0.7468 65 Sprat 100 logistic 3.1729 0.2637 0.4825 7.8 NA 8 Squid 100 betageneral 2.3843 1.2448 0.3454 1.9616 NA 4 Surimi 100 loglogistic 0.0747 1.2186 0.0014 2.8 0.0002 94 Trout 100 loglogistic 0.4389 1.7373 NA 11.8 0.4799 100 Tuna 100 lognormal 16.4224 2603.779 0.01 46.1106 -0.0018 50 Whiting 100 betageneral 0.3179 2.987 0.0301 0.6557 NA 16 Wolffish 100 loglogistic 1.1167 1.598 0.01 46.1106 -0.319 68


129

ANNEX III.10 - Distribution and parameters (Param) for the PCDD/F concentration in relevant seafood species, as well as the number of data points used (N) Species % Distribution Param1 Param2 Param3 Param4 Param5 N uniform min max betageneral α1 α2 min max normal µ σ Tr min Tr max lognormal µ σ Tr min Tr max shift logistic α β Tr min Tr max loglogistic β α Tr min Tr max γ (min) Anchovy 100 normal 0.1196 0.4232 0.0200 2.6600 NA 7 Anglerfish 100 loglogistic 0.0687 3.3976 0.0009 5.0000 -0.0162 106 Caviar 100 normal 0.5372 0.4404 0.0040 15.1400 NA 92 Cod, Iceland 15.46 betageneral 0.9217 2.5834 0.0018 0.0632 NA 17 Cod, North Sea 38.05 loglogistic 0.4288 69.2198 0.0150 0.6000 -0.3815 10 Cod, others 46.49 uniform NA NA 0.0151 0.1890 NA 4 Common shrimp 100 logistic 0.6012 0.2085 0.0500 3.1400 NA 9 Common whelk 100 normal 0.5529 0.4326 0.0095 7.2900 NA 95 Conger 100 normal 0.5372 0.4404 0.0040 15.1400 NA 92 Crab 100 uniform NA NA 0.0350 3.7500 NA 5 Eel 100 betageneral 0.8510 2.4160 0.1660 5.9656 NA 82 European plaice 100 loglogistic 0.7626 25.2105 0.0280 2.9600 -0.4768 19 Haddock 100 lognormal 0.0248 0.0350 NA 0.2000 0.0214 33 Halibut, Northeast Atlantic Ocean 42.43 logistic 0.6826 0.1760 0.0900 15.1400 NA 27 Halibut, Northwest Atlantic Ocean 57.57 uniform NA NA 0.0262 0.9190 NA 12 Herring, Baltic Sea 5.44 betageneral 0.4462 3.3479 0.7263 37.2970 NA 261 Herring, North Sea 50.89 loglogistic 1.3309 6.7805 0.0555 5.4400 -0.5048 63 Herring, others 43.67 normal 9.3960 6.8145 0.1970 40.0000 NA 4 John dory 100 normal 0.5529 0.4326 0.0095 7.2900 NA 95 Lobster 100 uniform NA NA 0.0625 1.1400 NA 3 Mackerel 100 loglogistic 0.3335 4.9583 0.0500 13.8200 -0.0225 63 Mussel 100 normal 0.4420 0.7677 0.0500 7.2900 NA 16 Nile perch 100 normal 0.5529 0.4326 0.0095 7.2900 NA 95 Norway lobster 100 normal 0.5529 0.4326 0.0095 7.2900 NA 95 Saithe and Pollack 100 normal 0.0573 0.0383 0.0009 0.3460 NA 17 Salmon, Baltic Sea 45.08 lognormal 3.7818 3.4750 NA 34.8000 1.4440 51 Salmon, others 54.92 loglogistic 0.4700 1.5254 NA 15.6000 0.1247 152 Sardine 100 lognormal 1.0494 0.4920 0.0100 2.4000 -0.5447 16 Scallop 100 loglogistic 0.0687 3.3976 0.0009 5.0000 -0.0162 106 Scampi 100 lognormal 4.7183 69167.9942 NA 0.5020 0.0320 15 Sea bream 100 normal 0.5372 0.4404 0.0040 15.1400 NA 92 Skate 100 loglogistic 0.0687 3.3976 0.0009 5.0000 -0.0162 106 Sole 100 normal 0.5529 0.4326 0.0095 7.2900 NA 3 Sprat 100 uniform NA NA 1.8595 4.1191 NA 23 Squid 100 betageneral 0.3239 1.8157 0.1069 1.7820 NA 9 Surimi 100 loglogistic 0.0687 3.3976 0.0009 5.0000 -0.0162 106 Trout 100 loglogistic 0.2700 1.6604 NA 9.6000 0.1293 99 Tuna 100 logistic 0.0462 0.6612 0.0040 10.0140 NA 42 Whiting 100 uniform NA NA 0.0177 0.0755 NA 14 Wolffish 100 logistic 0.5109 0.0190 0.1940 1.7100 NA 5


130

ANNEX III.11 - Distribution and parameters (Param) for total TEQ concentration in relevant seafood species, as well as the number of data points used (N)

Species % Distribution Param1 Param2 Param3 Param4 Param5 N uniform min max betageneral α1 α2 min max normal µ σ Tr min Tr max lognormal µ σ Tr min Tr max shift logistic α β Tr min Tr max loglogistic β α Tr min Tr max γ (min) Anchovy 100 loglogistic 1.6059 1.8653 0.015 49.0126 -0.6089 55 Anglerfish 100 loglogistic 0.1136 1.2063 NA 7.2 0.0232 88 Caviar 100 loglogistic 1.6059 1.8653 0.015 49.0126 -0.6089 55 Cod 100 betageneral 0.4188 1.6878 0.065 1.0239 NA 42 Common shrimp 100 normal 0.9698 0.6299 0.06 5.7 NA 7 Common whelk 100 normal 1.1112 0.8013 0.0139 8.12 NA 76 Conger 100 loglogistic 1.6059 1.8653 0.015 49.0126 -0.6089 55 Crab 100 loglogistic 0.1136 1.2063 NA 7.2 0.0232 88 Eel 100 loglogistic 7.888 2.2451 0.2295 252 -1.5753 106 European plaice 100 logistic 0.5841 0.0818 0.08 2.52 NA 17 Haddock 100 betageneral 0.3402 1.5712 0.0587 0.2269 NA 27 Halibut 100 normal 1.6462 0.8 0.35 38.46 NA 10 Herring, Baltic Sea 5.44 betageneral 0.8608 3.2638 1.3899 54.3135 NA 186 Herring, others 94.56 loglogistic 2.4942 2.2397 NA 59.4 0.6846 67 John dory 100 normal 1.1112 0.8013 0.0139 8.12 NA 76 Lobster 100 uniform NA NA 0.0646 1.845 NA 3 Mackerel 100 loglogistic 0.8483 3.1177 NA 55 0.5822 52 Mussel 100 uniform NA NA 0.12 6.09 NA 13 Nile perch 100 normal 1.1112 0.8013 0.0139 8.12 NA 76 Norway lobster 100 normal 1.1112 0.8013 0.0139 8.12 NA 76 Saithe and Pollack 100 normal 0.2928 0.1498 0.0036 1.8 NA 4 Salmon, Baltic Sea 31.32 loglogistic 17.8256 7.9011 1.485 65.96 -3.6945 50 Salmon, farmed 51.12 lognormal 6.8722 1.9378 0.05 30.8 -4.4695 176 Salmon, Pacific Ocean 17.56 betageneral 0.1903 1.1646 0.0999 0.5514 NA 8 Sardine 100 lognormal 16.4456 2.4166 0.05 15.8 -13.2736 6 Scallop 100 loglogistic 0.1136 1.2063 NA 7.2 0.0232 88 Scampi 100 betageneral 2.6814 4.9127 0.0381 0.163 NA 14 Sea bream 100 loglogistic 1.6059 1.8653 0.015 49.0126 -0.6089 55 Skate 100 loglogistic 0.1136 1.2063 NA 7.2 0.0232 88 Sole 100 normal 1.1112 0.8013 0.0139 8.12 NA 76 Sprat 100 normal 6.2662 1.1229 0.9233 15 NA 10 Squid 100 uniform NA NA 0.065 4.8 NA 4 Surimi 100 loglogistic 0.1136 1.2063 NA 7.2 0.0232 88 Trout 100 loglogistic 0.5654 1.5764 NA 21.2 0.6732 80 Tuna 100 betageneral 0.1073 1.4759 0.0283 33.7238 NA 41 Whiting 100 betageneral 0.1771 1.4633 0.077 0.9459 NA 15 Wolffish 100 loglogistic 1.6059 1.8653 0.015 49.0126 -0.6089 55

Chapter IV. Intake assessment results

131

Chapter IV. Probabilistic intake assessment of multiple compounds as a tool to quantify the nutritional-toxicological conflict related to seafood consumption


1. Sioen I, Van Camp J, Verdonck F, Verbeke W, Vanhonacker F, Willems J, De Henauw S.

Probabilistic intake assessment of multiple compounds as a tool to quantify the

nutritional-toxicological conflict related to seafood consumption. Chemosphere. Submitted.

2. Sioen I, De Henauw S, Verbeke W, Verdonck F, Willems J, Van Camp J. Fish

consumption is a safe solution to increase the intake of long chain omega-3 fatty acids.

Public Health Nutrition. Submitted.


132

1. Introduction

For the intake assessment of nutrients and contaminants via seafood consumption, it was

chosen to use a probabilistic approach, taking into account the variability of seafood

consumption data, body weight data, and nutrient and contaminant concentration data.

Therefore, a methodology and a software module were developed to calculate the

simultaneous intake of multiple (n=10) nutrients and contaminants via seafood

consumption, i.e. mercury (Hg), methyl mercury (MeHg; deduced from the Hg

concentration by a conversion factor), indicator PCBs (iPCBs), dioxin-like PCBs (dl PCBs),

dioxins plus furans (PCDD/Fs), total dioxin-like compounds (total TEQ), long-chain omega-

3 poly-unsaturated fatty acids (LC n-3 PUFAs), vitamin D, iodine, and fat. This software

module is applicable both for real intake assessments as well as for scenario analyses.

The results of both real intake assessments as scenario analyses related to seafood

consumption in Belgium are described in this chapter. The intakes were evaluated and

seafood consumption recommendations were formulated balancing the associated risks and

benefits to maximize public health. Firstly, the results of an intake assessment based on

existing seafood consumption data of the Belgian population are described. Within this part,

several scenarios related to the contamination data are considered. Secondly, intake

assessment results based on different consumption scenarios are described. These scenario

analyses were executed in order to find out whether the recommendation of LC n-3 PUFAs

can be reached without a contaminant intake of toxicological concern.


133

2. Intake assessment based on the current seafood consumption pattern in Belgium

In this part of chapter IV, the intake assessment based on the current seafood consumption

pattern (i.e. real intake assessment) is described. The results are discussed in part 4 of this

chapter.

2.1. Materials and Methods

2.1.1. Nutrient and contaminant data

The following contaminants were included: Hg and MeHg, seven indicator PCBs (iPCBs),

dioxin-like PCBs (dl PCBs), sum of 7 polychlorinated dibenzo-p-dioxin congeners (PCDDs)

and 10 polychlorinated dibenzofuran congeners (PCDFs) (from here on referred to as

PCDD/Fs), and total dioxin-like compounds (from here on referred to as total TEQ

(totTEQ)). Although several other contaminants may accumulate in the marine food chain

(e.g. arsenic, polycyclic aromatic hydrocarbons, polybrominated diphenyl ethers), it was

decided to limit this study to the above mentioned compounds. The main reason for this

limitation is the fact that for these contaminants a huge amount of data concerning different

species and different origins relevant for the Belgian market are available in international

scientific literature. Additionally, the rationale behind selecting the dioxin-like compounds

originates from the observation that seafood typically has a higher concentration of dioxin-

like compounds compared to other food items (when expressed per g fat) (Kiviranta et al,

2004). Moreover, a recent Belgian study showed that seafood is the most important dietary

source of dioxin-like compounds (see Chapter V part 1) (Bilau et al, 2007). The selection of

MeHg was motivated by the fact that seafood is the most important dietary source of Hg in

the human food chain. In the marine environment, inorganic Hg is to a high extent

transformed to MeHg, which further accumulates in the marine food chain. This organic

MeHg is for humans the most toxic form of Hg (Clarkson & Magos, 2006).

The following nutrients were included: the sum of EPA and DHA (EPA plus DHA), vitamin

D, iodine, and total fat. The rationale behind selecting those nutrients is that their natural

concentration in seafood is high compared to other food items. Fat was included specifically

to study the relation with the intake of the other compounds. In contrast to the other


134

nutrients, it was not relevant to discuss the intake of fat by seafood consumption only, since

a lot of other food items in the diet contribute to the total fat intake.

The concentrations and probability distributions used for both the nutrients and

contaminants originate from the two extensive, newly compiled databases described in

chapter III.

2.1.2. Consumption data

The intake assessment results described in this part of chapter IV were based on recorded

data of seafood consumption in Belgium. Two different existing food consumption

databases were used: one of Flemish adolescents and another of Belgian adults.

1. The first database contains consumption data of 341 adolescents (129 boys and 212 girls)

aged between 13 and 18 years, collected during seven consecutive days, using an

estimated food record method (semi-structured diary), in Ghent (Flanders, the Dutch-

speaking part of Belgium) between March and May 1997. It is the same database that was

used in chapter II to assess the overall PUFA and vitamin D intake of Flemish

adolescents.

2. The second database was collected from 821 Belgian adults (202 men and 619 women)

aged between 19 and 83 years, representative for the Belgian population with respect to

age and region. The data were collected between November and December 2004 as part

of the pan-European SEAFOODplus consumer survey (Honkanen & Brunsø, 2007). In

fact, a comprehensive questionnaire, including many behavioural and attitudinal aspects

relevant to consumer science was distributed to the study population. The questionnaire

assessed the frequency and the amount of consumption of the ten most consumed

seafood species in Belgium, i.e. cod, salmon, tuna, saithe, sole, European plaice, herring,

trout, mackerel, and eel. Methodological details of this study were recently described by

Olsen et al (2007) and Pieniak et al (2007).


135

2.1.3. Simulation model and probabilistic methodology

The following simulation model, combining species-specific seafood consumption data with

nutrient and contaminant concentration data, was used for the intake assessment:

i

avtivtvai bwT

CXY

⋅

⋅⋅= ∑ ∑∑ ) b( ,,,av,

where Yi= average daily intake of individual i per kg body weight (bw);

Xv,i,t= amount (g) of seafood species v consumed by individual i (with body weight

bwi), at day t (t = 1,…,T);

bv,a= probability determining whether seafood species v originates from region a;

Cv,a= concentration of a specific nutrient/contaminant in seafood species v from

region a.

As previously indicated, a probabilistic approach was applied for simulating, taking into

account the observed variability of the consumption, body weight, and concentration data.

Marien (2002) previously described the importance of applying species-specific consumption

data on individual level as well as individual body weight data when assessing intake of

contaminants through seafood. To perform the simulations of this probabilistic intake

assessment, a software module called ProbIntakeUG was developed at the Ghent University

(Ghent, Belgium). ProbIntakeUG is applicable in the free available software program R® (R

Development Core Team, 2006). For the consumption, the variability was taken into account

in a non-parametric way, i.e. by using the data as such, because the dataset was large

enough. For the concentration data, the variability was taken into account in a parametric

way, i.e. by using species and compounds specific probability distributions. For the body

weight data, a non-parametric approach was applied in the simulations using the adolescent

consumption data, since the body weight of each individual adolescent was known and as

such, no assumptions needed to be made. In contrast, for the adults’ consumption dataset no

individual body weight data were available. Therefore, normal body weight probability

distribution were applied per gender and age interval, based on available data for the

Belgian population (B.I.R.N.H. study (De Backer, 1984; Kornitzer & Dramaix, 1989)) (Table

IV.1).


136

Table IV.1: Mean and standard deviation (S.D.) of the applied body weight distributions * Age interval (years) Men Women Mean (kg) S.D. Mean (kg) S.D. 30-39 77.2 11.2 62.7 10.9 40-49 78.9 11.5 66.7 11.7 50-59 77.4 11.4 69.5 11.2 60-69 75.3 12.3 69.5 11.9 * From references (De Backer, 1984; Kornitzer & Dramaix, 1989)

The simulation procedure for each individual works as follows: each single consumption

data point is multiplied with a concentration data point. This combination is conducted for

all consumed seafood species and for all different compounds. Next, the assessed intakes per

compound were enumerated and this sum was divided by the number of days and the

individual’s body weight. Finally, this procedure was repeated for all individuals.

Two extensions were applied increasing the number of individuals and the number of

recording days per individual:

- First of all, the number of individuals was artificially repeated ten times. This repetition

accounted predominantly for the uncertainty arising from a relatively small sample size

when compared to the total population.

- Secondly, the 7-day diary was extended to a fictitious 35-day diary per adolescent (by

simply repeating the diary five times consecutively) and the adults’ consumption data

were extended to ten weeks per adult (by repeating the data ten times), accounting for

the uncertainty arising from a relatively short period in time.

Repeating the dataset was needed to ensure that the whole concentration range of the

nutrients and contaminants was reflected in the intake calculation and that a good

convergence of the population intake assessment was reached. A larger extension would

increase the simulation processing time, without significantly improving the intake

estimations.

2.1.4. Scenario analyses related to the contamination level

The probabilistic intake assessment based on real consumption data was run for three

scenarios, differing on the level of contaminant concentrations used:


137

1. In a first scenario, all the concentration data in the contaminant database were used,

whether or not the EU maximum levels for seafood products from the European

Commission (2006) were exceeded.

2. In a second scenario, the intake of dl PCBs, PCDD/Fs, and total TEQ has been estimated

excluding concentration data from the herring and salmon originating from the Baltic

Sea, hence using only the concentration data from herring and salmon originating from

other fishing regions. The rationale behind this scenario is that the Baltic Sea has been

contaminated for many years by emissions of PCDD/Fs from paper and metal industry

plants and waste incineration plants, as well as from rivers discharging into the Baltic,

leading to higher concentrations in seafood (Danish Veterinary and Food Administration,

2004). This can also be seen in the box plots for dioxin-like compounds given in chapter

III. In order to reduce human consumption of dioxin-like compounds, in July 2002 the

European Commission set a new maximum allowable concentration of PCDD/Fs in

edible parts of seafood of 4 pg WHO-TEQ/g fresh weight. Finland and Sweden got an

exemption order until the end of 2006 to place seafood from the Baltic region with

PCDD/F levels above this maximum allowable concentration on the domestic market.

3. In a third scenario, the intake of dioxin-like compounds and total TEQ has been

estimated excluding all concentration data exceeding the EU maximum levels, being 4 pg

WHO-TEQ/g fresh weight for PCDD/Fs and 8 pg WHO-TEQ/g fresh weight for dl

PCBs plus PCDD/Fs (total TEQ), with the exception of eel that may contain 12 pg WHO-

TEQ/g fresh weight of dl PCBs plus PCDD/Fs (European Commission, 2006). The

rationale behind this third scenario was to assess the intake of dioxin-like compounds

under the assumption that the EU regulation is applied under the strictest conditions.

2.1.5. Evaluation of nutrient and contaminant intakes

To evaluate whether the intake of nutrients was adequate, ‘ad hoc’ reference values were

calculated for the two study populations on the basis of the dietary reference intakes (DRI)

formulated by the Belgian Health Council (2007).

- For EPA plus DHA, this DRI is 0.3% of the total energy intake per day. For the

adolescents study population, the reference value for EPA plus DHA was calculated for

each individual based on his/her energy intake and body weight; the mean EPA plus

DHA reference value for this population was 12.2 mg/kg bw/day (Matthys et al, 2003).


138

For the Belgian adult study population, a mean bw of 70 kg was applied and a mean

energy intake of 2046 kcal was used, based on the data of the most recent Belgian Food

Consumption Survey (De Vriese et al, 2006), leading to a reference value of 9.7 mg/kg

bw/day for EPA plus DHA.

- For vitamin D and iodine, the Belgian DRI is 5 µg/day (applied in this approach as 0.085

µg/kg bw/day for the adolescents (based on a mean bw of 59.1 kg) and 0.071 µg/kg

bw/day for the adults (based on a mean bw of 70 kg)) and 150 µg/day (applied in this

approach as 2.54 µg/kg bw/day for the adolescents and 2.14 µg/kg bw/day for the

adults), respectively. Dividing the reference values for nutrients by the bw was relevant

in this study in order to express the reference values for nutrients and contaminants on

the same scale.

Tolerable daily/weekly intakes (TDI/TWI) were used to evaluate contaminant intakes. For

MeHg, a TWI of 1.6 µg/kg of bw/week (i.e. 0.228 µg/kg bw/day) has been proposed (EFSA,

2004) and for dioxin-like compounds, the EU proposes 2 pg WHO-TEQ/kg bw/day (14 pg

WHO-TEQ/kg bw/week) (Scientific Committee on Food, 2001). For the non-dioxin like

PCBs (ndl PCBs), no health based guidance value for humans has been established. The

health based advice is, therefore, that the intake of ndl PCBs should be as low as possible

(EFSA, 2005c). Finally, it is important to note that the thresholds used for nutrient and

contaminant intake refer to the intake via the total diet, whereas in this study, they were

used to evaluate the intake via one single group of food items, namely seafood.

2.2. Results

Of all 341 adolescents in the first consumption database, 123 (36%) did not consume any

seafood during the week of the study; hence, their assessed multiple compound intake via

seafood was equal to zero. Of all 821 respondents in the database of Belgian adults, 52 (6.3%)

claimed not to consume seafood. The results in the tables are provided for both the whole

population sample and the consumers-only (Table IV.2 and Table IV.3).

The assessed contaminant intake results are given in Table IV.2.

Table IV.2: Summary of assessed intakes of contaminants via seafood consumption for Flemish adolescents and Belgian adults, for the whole population sample and for the

seafood consumers only (the intakes exceeding the TDI are indicated in bold)

Including all concentration data

Excluding concentration data of Baltic

herring and salmon

Excluding concentrations data

> EU max levels

MeHg iPCB dl PCB PCDD/Fs totTEQ dl PCB PCDD/Fs totTEQ PCDD/Fs totTEQ

ng/kg bw/day pg WHO-TEQ/kg bw/day pg WHO-TEQ/kg bw/day pg WHO-TEQ/kg bw/day

All Cons. All Cons. All Cons. All Cons. All Cons. Cons. Cons. Cons. Cons. Cons.

TDI 228 - 2 2 2 2 2 2 2 2

Mean 16.8 26.3 3.1 4.8 0.4 0.6 0.2 0.4 0.5 0.8 0.4 0.2 0.5 0.2 0.4

P50 5.3 16.4 0.4 1.5 0.0 0.2 0.0 0.1 0.1 0.3 0.1 0.1 0.2 0.1 0.2

P90 47.3 64.3 5.9 7.6 1.1 1.6 0.7 1.0 1.5 2.2 0.9 0.6 1.4 0.6 1.1

P95 73.3 90.7 8.5 10.3 1.8 2.3 1.3 1.6 2.6 3.2 1.3 0.9 2.0 0.8 1.8

P97.5 97.5 116.1 11.7 14.3 2.7 3.3 1.8 2.2 3.4 4.0 1.7 1.4 2.8 1.1 2.4

AD

OLE

SCEN

TS

P99 133.1 150.4 18.1 21.7 3.7 4.0 2.5 2.9 4.8 5.4 3.0 2.1 3.5 1.4 3.1

Mean 42.7 45.6 6.4 6.8 0.8 0.9 0.5 0.5 0.9 1.0 0.6 0.4 0.7 0.3 0.5

P50 28.70 31.2 3.1 3.4 0.5 0.6 0.3 0.4 0.6 0.7 0.4 0.2 0.5 0.2 0.3

P90 91.8 94.7 13.8 14.5 1.9 1.9 1.1 1.1 2.1 2.3 1.2 0.8 1.5 0.6 1.0

P95 125.3 128.6 22.6 23.5 2.5 2.6 1.5 1.5 2.9 3.0 1.7 1.1 2.0 0.8 1.4

P97.5 164.9 167.6 35.0 36.5 3.2 3.2 1.9 1.9 3.5 3.7 2.3 1.4 2.5 1.0 1.7

AD

ULT

S

P99 229.1 232.3 56.1 58.0 4.6 4.7 2.6 2.6 4.9 5.0 3.1 1.9 3.2 1.3 2.2

Cons.: consumers-only; TDI: tolerable daily intake; for the choice of the TDI see text.


140

The intake assessment of MeHg via seafood for both populations indicates that only a small

percentage of the studied populations (~ 1%) exceeded the MeHg TDI through seafood

consumption. A high correlation was found between the amount of tuna consumed and the

intake of MeHg (r²=0.87 for the total group of adult seafood consumers and r²=0.90 for adult

tuna consumers). The iPCBs intake via seafood was characterized by some extreme intakes,

as illustrated by the 99th percentiles (Table IV.2). A high correlation was found between the

amount of eel consumed and the iPCB intake (r²=0.82 for the total group of adult seafood

consumers and r²=0.90 for adult eel consumers). The intake of dioxin-like compounds was

calculated for the three different scenarios, for both populations. It is important to note that

the intake assessment for dl PCBs, PCDD/Fs, and total TEQ all start from independent data

sets. There is, hence, no direct link between the TEQ dl PCB, the TEQ PCDD/F, and the total

TEQ intake (Table IV.2).

Table IV.3 gives the results of the assessed nutrient intakes for the adolescents and the

adults, respectively, showing that only a very small percentage of the studied populations

achieved the reference values via seafood consumption.

Table IV.3: Summary of the intake assessment for nutrients via seafood consumption for Flemish adolescents and Belgian adults, for the whole population sample and for the seafood consumers only (the figures given in bold are intakes higher than the reference value for that nutrient) EPA and DHA Vitamin D Iodine (mg/kg bw/day) (µg/kg bw/day) (µg/kg bw/day) All Cons. All Cons. All Cons.

Reference value 12.2 0.085 2.54 Mean 1.7 2.73 0.010 0.015 0.28 0.44 50th percentile 0.6 1.56 0.001 0.005 0.06 0.20 90th percentile 5.3 7.59 0.030 0.046 0.94 1.19 95th percentile 8.4 10.03 0.052 0.064 1.32 1.46 97.5th percentile 10.7 12.05 0.069 0.085 1.52 1.74 A

DO

LESC

ENTS

99th percentile 14.0 14.61 0.096 0.110 2.07 2.43 Reference value 9.7 0.071 2.14 Mean 3.5 3.8 0.023 0.024 0.33 0.36 50th percentile 2.6 2.8 0.016 0.017 0.25 0.27 90th percentile 7.7 7.9 0.050 0.051 0.72 0.75 95th percentile 10.1 10.4 0.067 0.067 1.00 1.02 97.5th percentile 13.1 13.6 0.086 0.087 1.29 1.31

AD

ULT

S

99th percentile 18.0 18.2 0.119 0.122 1.54 1.55 Cons.: consumers-only; The reference values are based on the dietary reference intakes proposed by the Belgian Health Council (2007), but are expressed in function of body weight, for explanation see text


141

Fig. IV.1 presents a two-dimensional matrix of results for the nutrients and contaminants

under study (after a log-transformation) based on the consumption database of the adults

and on contamination scenario 1 (taking into account all concentration data).

Fig. IV.1 Matrix visualising the log-transformed results of the probabilistic intake assessment via seafood

consumption for the ten compounds of interest for Belgian adults; mercury (Hg), methyl mercury (MeHg), sum of 7 indicator PCBs (iPCBs), dioxin-like PCBs (dl PCBs), PCDD/Fs (dioxins), total dioxin-like compounds

(totTEQ), EPA and DHA, vitamin D (vitD), iodine, and fat

On the main descending diagonal axis of the matrix, frequency distributions of the intake

assessment results of all individual compounds are shown. At the lower left half of the

matrix, 45 scatter plots show the combined intake for each possible pair of nutrients and/or

contaminants. At the upper right, the correlation coefficients between the intakes of each pair

of compounds are given. To facilitate the interpretation, the font size is related to the size of

the correlation coefficient. The highest correlations are observed between the assessed intake

of several fat-soluble compounds, e.g. r²(EPA plus DHA and vitamin D) = 0.94, r²(total TEQ

and EPA plus DHA) = 0.87. A negligible correlation exists between the iodine and (Me)Hg

intake and all the other compounds, respectively. The figure shows relatively high

correlations between (non) dioxin-like PCBs and PCDD/Fs and EPA plus DHA. This


142

indicates that it is hard to achieve a higher EPA plus DHA intake via seafood consumption

without increasing the intake of ndl PCBs and dioxin-like compounds.

Based on the adolescent and adult food consumption databases, an evaluation of the intake

of LC n-3 PUFAs (EPA plus DHA) and dioxin-like compounds (total TEQ) can be combined

in the output of the ProbIntakeUG module. Fig. IV.2 (A1 and A2) provides a scatter plot,

focussing on EPA plus DHA and total TEQ, after exclusion of the concentration data of Baltic

herring and salmon.

1.E-04

1.E-03

1.E-02

1.E-01

1.E+00

1.E+01

1.E-04 1.E-03 1.E-02 1.E-01 1.E+00 1.E+01 1.E+02

totTEQ (intake/TDI)

EPA

&D

HA

(int

ake/

RD

A)

43

1 2

1.E-03

1.E-02

1.E-01

1.E+00

1.E+01

1.E-03 1.E-02 1.E-01 1.E+00 1.E+01

totTEQ (intake/TDI)

EPA

&D

HA

(int

ake/

RD

A)

43

1 2

1.E-04

1.E-03

1.E-02

1.E-01

1.E+00

1.E+01

1.E-04 1.E-03 1.E-02 1.E-01 1.E+00 1.E+01 1.E+02

totTEQ (intake/TDI)

EPA

&D

HA

(int

ake/

RD

A)

43

21

1.E-03

1.E-02

1.E-01

1.E+00

1.E+01

1.E-03 1.E-02 1.E-01 1.E+00 1.E+01 1.E+02

totTEQ (intake/TDI)

EPA

&D

HA

(int

ake/

RD

A)

43

1 2

Fig. IV.2 Ratio of intake of total TEQ divided by TDI (2 pg WHO-TEQ/kg bw/day) and intake of EPA plus DHA divided by DRI (0.3% of total energy requirement) as a result of the current seafood consumption of Belgian adolescents (A1) and adults (A2) and as a result of doubling of the current seafood consumption of Belgian adolescents (B1) and adults (B2) with exclusion of concentration data of Baltic seafood (log-scale)

The graphs in Fig. IV.2 show the ratio of the intake of total TEQ (pg WHO-TEQ/kg bw/day)

divided by the TDI (2 pg WHO-TEQ/kg bw/day) and the intake of EPA plus DHA (mg/kg

bw/day) divided by the reference value. By expressing the intakes relative to the reference

values, the discriminative value for being at the ‘risk’ versus at the ‘benefit’ side is ‘1’ on both

axes. As such, four quadrants (zones) are obtained, all with a relevant interpretation

describing whether or not a sufficient amount of seafood was consumed to meet the EPA

plus DHA recommendation, and with or without exceeding the 2 pg WHO-TEQ/kg bw/day

limit. Four quadrants (zones) are obtained:

A2 A1

B1 B2


143

- Zone 1: consumption of sufficient amount of seafood to meet the EPA plus DHA DRI,

without exceeding the 2 pg WHO-TEQ/kg bw/day limit (most beneficial zone);

- Zone 2: consumption of sufficient amount of seafood to meet the EPA plus DHA DRI, but

exceeding the 2 pg WHO-TEQ/kg bw/day limit;

- Zone 3: consumption of too little seafood to meet the EPA plus DHA DRI, and not

exceeding the 2 pg WHO-TEQ/kg bw/day limit;

- Zone 4: consumption of too little seafood to meet the EPA plus DHA DRI, but exceeding

the 2 pg WHO-TEQ/kg bw/day limit (worst case zone).

Only a few individuals met their EPA plus DHA requirement without exceeding the limit for

total dioxin-like compounds. It was useful to assess the effect of doubling the seafood

without any other changes to their dietary pattern (no alterations of the species consumed),

shown in Fig. IV.2 (B1 and B2). The point cloud representing the EPA plus DHA on total

TEQ intake ratio shifted to the upper right corner (towards zone 2), indicating that more

individuals met their EPA plus DHA intake, but at the same time increased their intake of

dioxin-like compounds proportionally.


144

3. Intake assessment based on consumption scenarios

In this third part of chapter IV, the intake assessment based on the seafood consumption

scenarios is described. The results are discussed in part 4 of this chapter together with the

results of the intake assessment based on real seafood consumption data.

3.1. Materials and Methods

The consumption scenario analyses started from the current Belgian seafood species

consumption pattern and the recommendation to consume seafood twice a week. Different

variations were elaborated on (1) the fish species consumed and (2) the frequency of fish

consumption:

- First, the seafood species consumption pattern was artificially changed in two different

ways:

o people increasing their fatty fish consumption up to 50% of their total fish

consumption (simulated pattern n°1)

o people replacing all the lean fish species by fatty fish species (simulated pattern

n°2).

- Second, in relation to the frequency of seafood consumption, two variations on the

consumption recommendation were considered:

o people consuming seafood only once a week (50% of recommendation)

o people consuming seafood three times a week (150% of recommendation).

A scheme of the different scenarios is presented in Fig. IV.3. The different aspects are

described in detail below.


145

Fig. IV.3 Scheme of the elaboration and implementation of the different scenarios

3.1.1. Nutrient and contaminant data

The same probability distributions of nutrient and contaminant concentrations are used as

described in chapter III. However, for this consumption scenario study, only the sum of EPA

and DHA concentrations was considered; the other nutrients were not taken into account.

Reason for this is that the aim of the consumption scenario analyses was to find out whether

the recommendation of LC n-3 PUFAs can be reached without a contaminant intake of

toxicological concern. Concerning the contaminants, this consumption scenario analyses

focused MeHg, dl PCBs, PCDD/Fs, and total dioxin-like compounds (total TEQ) because


146

toxicological concern was most relevant for these four compounds. Considering the

concentrations of dioxin-like compounds in salmon and herring, contaminant concentrations

measured in Baltic salmon and herring were excluded from the analyses executed in this

scenario study, as their presence on the Belgian market is considered as negligible due to a

European regulation.

3.1.2. Consumption and body weight data

Three times three different consumption scenarios were investigated, starting from the

current seafood species consumption pattern taking into account the seven most consumed

seafood species (Table IV.4), determined through the pan-European SEAFOODplus

consumer survey (Honkanen & Brunsø, 2007). No crustacean or molluscs species were taken

into account in the scenarios, since all of the seven most consumed seafood species were

finfish species. Two hypothetical scenarios with an altered pattern of fish species

consumption were constructed: (1) assuming that the population increases the contribution

of fatty fish (> 5% fat) consumption to 50% of the total seafood consumption, and (2)

considering that people replace all lean fish species (≤ 5% fat) by fatty fish species. The

contribution of the different species in both scenarios was calculated proportionally to their

current contribution. For the three different consumption patterns considered, three sub

scenarios were investigated differing in consumption frequency, i.e. once, twice, or three

times a portion of 150 g seafood per week.

Table IV.4: Contribution of the seven different seafood species (%) to the total seafood consumption for three consumption scenarios, as well as the ratio of the concentration of EPA plus DHA on methyl mercury (MeHg) and total dioxin-like compounds (totTEQ) Species Percentage of total seafood consumption (%) Concentration ratio (median,

[5th percentile, 95th percentile])

Current consumption

pattern 50% lean & 50

% fatty fish Only fatty

fish EPA&DHA / MeHg (10-3)

EPA&DHA / totTEQ (10-6)

Cod 24.7 19.0 0.0 0.06 [0.03-0.11] 18.89 [4.04-59.13] Tuna 19.2 14.7 0.0 0.02 [0.00-0.12] 2.39 [0.19-42.38] Alaska Pollack 13.6 10.6 0.0 0.11 [0.05-0.26] 13.75 [5.09-58.74] Plaice 7.4 5.7 0.0 0.16 [0.04-0.36] 5.57 [1.39-13.98] Total lean fish 64.9 50.0 0.0 Atlantic Salmon 19.7 28.0 56.0 0.93 [0.31-1.99] 8.73 [2.35-57.64] Herring 8.0 11.4 22.8 0.45 [0.20-2.06] 3.92 [1.20-9.86] Mackerel 7.4 10.6 21.2 0.94 [0.31-2.28] 10.33 [5.03-23.05] Total fatty fish 35.1 50.0 100.0


147

For the intake assessment, a population of 600 individuals was used (300 men - 300 women),

equally divided over four different age classes (30-39y; 40-49y; 50-59y; 60-69y). Normal body

weight distributions were applied per gender and age interval, based on available data for

the Belgian population, being the same as applied for the real intake assessments based on

the adults’ consumption data described in the previous part of this chapter (B.I.R.N.H study

(De Backer, 1984; Kornitzer & Dramaix, 1989)) (Table IV.1, given in the previous part of this

chapter). A number of 600 individuals was sufficient ensure that the whole concentration

range of the nutrients and contaminants was reflected in the intake calculation and that a

good convergence of the population intake estimated was reached.

3.1.3. Simulation model and probabilistic methodology

The simulation model, the approach, and the software module (ProbIntakeUG) used for the

consumption scenario analyses are equal as described in part 2.1.3. of this chapter. The sole

difference is that it was assumed that consumers kept this consumption pattern for a whole

year (52 weeks) for the purpose of optimising integration of the inter-species variability in

the nutrient and contaminant concentrations during the intake assessment and to calculate at

the end the average daily intake over a long term period.

3.1.4. Evaluation of nutrient and contaminant intakes

Similar to what was done for the intake assessment based on real consumption data,

reference values were used to evaluate nutrient and contaminant intakes. To evaluate

population intakes of EPA plus DHA, an ‘ad hoc’ reference value of 681 mg/day or 9.7

mg/kg bw/day for EPA plus DHA was calculated starting from the existing Belgian DRI

equal to 0.3% of the total energy intake (Belgian Health Council, 2007) and assuming a mean

body weight of 70 kg and a mean energy intake of 2046 kcal, the latter based on the data of

the most recent Belgian Food Consumption Survey (3245 individuals above 15 years; 1623

women, 1622 men) (De Vriese et al, 2006). Dividing the EPA plus DHA reference value by the

body weight was relevant in this study in order to express the reference values for nutrients

and contaminants on the same scale. For MeHg, a tolerable weekly intake (TWI) of 1.6 µg/kg

of bw/week (i.e. 0.228 µg/kg bw/day) is proposed (EFSA, 2004) and for dioxin-like

compounds, the EU proposes 2 pg WHO-TEQ/kg bw/day (Scientific Committee on Food,

2001).


148

3.1.5. Inclusion of LC n-3 PUFA enriched margarine

Currently, EPA and DHA enriched margarine is commonly available on the Belgian market

and is therefore considered in this study. The EPA and DHA concentration in enriched

margarine varies a lot depending on the brand, but it varies also in time. A first brand of

margarine available in Belgium claims that their EPA plus DHA enriched margarine contains

5 mg EPA plus DHA/g margarine. A second brand indicated that 1 g of enriched margarine

contains 7.5 mg EPA plus DHA. A third supplier stated that its enriched variant of

margarine contains 0.9 mg DHA/g margarine. Belgian dieticians assessed that one slice of

bread with a regular layer of margarine contains 5 g of margarine (Belgian Health Council,

2005). Assuming a consumption of 4 to 7 slices of bread a day leads to a consumption of 20 to

35 g of margarine a day and 100 to 262.5 mg EPA plus DHA per day (using the two EPA plus

DHA-richest versions of enriched margarine). The results of the most recent Belgian Food

Consumption Survey (De Vriese et al, 2006) indicated that currently the mean daily

consumption of culinary fats and margarines is 21.2 g with an interquartile range of 6.0 g to

28.6 g. In the scenario analyses executed, it was assumed that all consumers would use daily

the average amount of enriched margarine containing 7.5 mg EPA plus DHA/g margarine.

3.2. Results

Table IV.4 (given previously in this part of the chapter) shows that currently 65% of the total

seafood consumption in Belgium is composed of lean fish species (≤ 5% fat), with cod being

the most important species. Salmon is the most consumed fatty fish (> 5% fat) in Belgium.

Table IV.4 also shows the median, the 5th and the 95th percentile of the species specific ratio of

the EPA plus DHA concentration on MeHg and total TEQ concentration. The higher the

ratio, the higher the nutrient concentration compared to the contaminant concentration. The

results illustrate that for some species the distribution of the ratio is very wide and skewed to

the right, e.g. EPA&DHA/totTEQ for tuna and salmon (Table IV.4).


149

3.2.1. Seafood as only source of EPA and DHA

Table IV.5, Fig. IV.4, and Fig. IV.5 show the intake assessment results for the different

scenarios.

Table IV.5: Mean intake of the different compounds for three different seafood consumption patterns and three different scenarios of consumption frequency MeHg iPCB dl PCB PCDD/F totTEQ EPAplusDHA ng/kg bw/day pg WHO-TEQ/kg bw/day mg/kg bw/day 1x 150 g per week Current pattern 36.19 2.63 0.33 0.24 0.40 2.66 50% lean and 50% fatty 30.40 2.88 0.39 0.30 0.41 3.22 Only fatty fish 8.28 3.44 0.54 0.51 0.58 5.14 2x 150 g per week Current pattern 70.01 5.26 0.68 0.52 0.79 5.32 50% lean and 50% fatty 57.93 5.70 0.77 0.64 0.82 6.43 Only fatty fish 16.52 6.86 1.08 1.08 1.14 10.28 3x 150 g per week Current pattern 106.19 7.98 1.02 0.73 1.20 8.01 50% lean and 50% fatty 86.45 8.50 1.14 0.94 1.23 9.64 Only fatty fish 24.79 10.27 1.60 1.56 1.74 15.41 MeHg = methyl mercury; iPCB = seven indicator PCBs; dl PCB = dioxin-like PCBs; totTEQ = total dioxin-like compounds

The results indicate that changing the current seafood consumption pattern by increasing the

contribution of fatty fish will reduce the intake of MeHg (Table IV.5). This could already be

concluded based on the comparison between the EPA plus DHA over MeHg ratio of lean

and fatty seafood species (Table IV.4). Nevertheless, in none of the scenarios considered,

consumers exceed the TDI for MeHg on a long term base. In contrast to MeHg, the intake of

the other contaminants increases when replacing lean fish species by fatty fishes. This was

expected given the lipophilic character of these contaminants. Simultaneously, increasing the

contribution of fatty fish species increases the intake of beneficial EPA plus DHA. Some lean

species also have a relative high EPA plus DHA over total TEQ ratio compared to other

species, e.g. cod and Pollack (Table IV.4), but the absolute EPA plus DHA concentration in

these species is so low that an unrealistically large amount of these species should be eaten to

achieve the recommended EPA plus DHA intake.


150

1E-1 1E04E-2 5E-2 7E-2 2E-1 3E-1 4E-1 5E-1 7E-1

1E-1

1E0

2E-1

3E-1

4E-1

5E-1

6E-1

7E-18E-19E-1

2E0

1x150g

Current consumption pattern50% lean and 50% fatty fishOnly fatty fish

MeHg (intake/TDI)

EP

A&

DH

A (i

ntak

e/D

RI)

Fig. IV.4 Methyl mercury (MeHg) intake divided by the TDI (228 ng/kg bw/day) in relation to the EPA plus

DHA intake divided by the DRI (9.7 mg/kg bw/day) for three different seafood consumption patterns and three different scenarios of consumption frequency (logarithmic scales); 1x, 2x, 3x seafood consumption per week:

clouds from lower left to upper right; the limit value for being at risk due to a too high MeHg intake or inadequate EPA plus DHA intake is ‘1’ on both axes; an extra reference line was added at half of the reference

value for EPA plus DHA

Two scatter plots are provided, focussing on EPA plus DHA and MeHg (Fig. IV.4) and total

TEQ (Fig. IV.5), based on the results of the different consumption scenarios. The plots show

the intake of MeHg and total TEQ, respectively, divided by their TDI in relation to the intake

of EPA plus DHA divided by the reference value (9.7 mg/kg bw/day). Consequently, the

limit value for being at risk due to a too high contaminant intake or inadequate EPA plus

DHA intake is ‘1’ on both axes. On both scatter plots, extra reference lines were added: (1) at

half of the TDI for total TEQ, to take into account that the human diet contains other sources

of dioxin-like compounds besides seafood; (2) at half of the reference value for EPA plus

DHA, since the Belgian DRI for EPA plus DHA is high compared to other countries (see

discussion part of this chapter). By adding these reference lines, different zones are obtained,

all with a relevant interpretation describing whether or not a sufficient amount of seafood

was consumed to meet the DRI for EPA plus DHA, with or without exceeding the

contaminant TDI.

2x150g

3x150g


151

1E-1 1E02E-1 3E-1 4E-1 5E-1 6E-1 8E-1 2E0

1E-1

1E0

2E-1

3E-1

4E-1

5E-1

6E-1

7E-18E-19E-1

2E0

Current consumption pattern50% lean and 50% fatty fishOnly fatty fish

TotTEQ (intake/TDI)

EP

A&

DH

A (i

ntak

e/D

RI)

Fig. IV.5 Intake of total dioxin-like compounds (totTEQ) divided by the TDI (2 pg WHO-TEQ/kg bw/day) in

relation to the EPA plus DHA intake divided by the DRI (9.7 mg/kg bw/day) for three different seafood consumption patterns and three different scenarios of consumption frequency (logarithmic scales); 1x, 2x, 3x

seafood consumption per week: clouds from lower left to upper right; the limit value for being at risk due to a too high totTEQ intake or inadequate EPA plus DHA intake is ‘1’ on both axes; two extra reference lines were added (1) at half of the TDI for total TEQ, to take into account that the human diet contains other sources of dioxin-like

compounds besides seafood; (2) at half of the reference value for EPA plus DHA

Considering the EPA plus DHA intake, the results show that only a seafood consumption

pattern consisting for 50% lean fish species and 50% fatty fish species with a minimum

consumption frequency of three times a week, or a seafood consumption pattern consisting

only of fatty species with a frequency of minimum twice a week, will lead to an adequate

EPA plus DHA intake using the Belgian DRI and not taking into account other sources of

these fatty acids. Fig. IV.4 shows that none of the considered consumption scenarios will lead

to exceeding of the TDI for MeHg, indicating that the mercury contamination of seafood

available on the Belgian market is not an issue of major concern. In contrast, when

consuming three times a week a portion of fatty fish, the intake of dioxin-like compounds

will approach the TDI value (Fig. IV.5). Knowing that the human diet contains also other

important sources of dioxin-like compounds, an intake of three portions of fatty fish per

week may be of toxicological concern.


152

3.2.2. Enriched margarine as extra dietary sources of EPA and DHA

Assuming that all consumers would use daily 21.2 g enriched margarine containing 7.5 mg

EPA plus DHA/g margarine, this would lead to a mean daily intake of 159 mg EPA plus

DHA, being 23.3% of the Belgian DRI (681 mg/day). In Fig. IV.6, scatter plots are shown for

the different seafood consumption scenarios with and without adding enriched margarine as

an LC n-3 PUFA source (mean daily intake of 21.1 g), assuming that the margarine

consumption will not contribute to the intake of contaminants. Consuming enriched

margarine will help to increase the EPA plus DHA intake. Nevertheless, the contribution is

rather limited and margarine as such is not sufficient to reach the DRI. A consumption

scenario of weekly 150 g lean fish and 150 g fatty fish combined with a daily consumption of

LC n-3 PUFA enriched margarine leads to an EPA plus DHA intake close to the DRI with a

mean total TEQ intake below half of the TDI.

1E-1 1E0totTEQ (intake/TDI)1E-1

1E0

50% lean and 50% fatty fish

EPA

&DH

A (in

take

/DR

I)

Only fatty fish1E-1 1E0totTEQ (intake/TDI)

1E-1

1E0

Current consumption pattern

EPA&

DH

A (in

take

/DR

I)

Seafood onlySeafood and omega-3 enriched margarine

1E-1 1E0totTEQ (intake/TDI)

1E0

EPA

&D

HA

(inta

ke/D

RI)

Fig. IV.6 Intake of total dioxin-like compounds (totTEQ) divided by the TDI (2 pg WHO-TEQ/kg bw/day) in

relation to the EPA plus DHA intake divided by the DRI (9.7 mg/kg bw/day) for three different seafood consumption patterns and three different scenarios of consumption frequency, with and without taking omega-3

enriched margarine into account (logarithmic scales); 1x, 2x, 3x seafood consumption per week: clouds from lower left to upper right


153

4. Discussion

4.1. Intake assessment of multiple compounds via seafood consumption

4.1.1. Results of the intake assessment based on real seafood consumption data

Related to consumption datasets used in the first part of this chapter, some limitations have

to be mentioned. First, it is relevant to reiterate that the two different seafood consumption

databases have been collected by two different methodologies, i.e. a food record indicating

all consumed seafood products versus a food frequency questionnaire asking consumption

information on the ten most consumed seafood species, and in two different periods (1997

and 2004). In both food consumption databases, seasonal variation in seafood consumption

was not taken into account. Nevertheless, these were the only consumption data available for

the Belgian population giving sufficiently detailed information on the weekly intake of

seafood consumption.

Another shortcoming is that the consumption data of the adolescents describe a short-term

intake, whereas it would be more appropriate to end up with long term intake data (i.e.

usual intake) to compare with a TDI. Consequently, the assessed variance of the intakes

includes not only between-person variability, but also within-person variability, while the

latter is of less interest for long term intake assessments. In addition, it was assumed in this

paper that non-seafood consumers in the consumption studies were people who did never

consume any seafood. Both assumptions lead to an overestimation of the assessed variance

of the nutrient and contaminant intake distributions. The overestimation due to within-

person variability is smaller when dietary records are used during a longer period in time (in

this study, a seven day dietary record was used for the adolescent food consumption

database). Statistical modelling techniques exist filtering within-person variability to

calculate usual intake starting from short term consumption data and taking into account

that non-consumers can be consumers on other moments in time and vice versa (Dodd et al,

2006; Nusser et al, 1996; Slob, 1993; Tooze et al, 2006). These methods were not used here as

the improvement in accuracy was considered to be small relative to other sources of

uncertainty in this assessment, e.g. uncertainty related to the contamination data due to

different analytical methods, different sampling plans, etc.


154

The results of the intake assessment based on real seafood consumption data show that the

seafood consumption of the considered populations is not sufficient to reach the DRI for

EPA plus DHA, vitamin D, and iodine. The current seafood consumption does not lead to a

MeHg intake of major toxicological concern. In contrast, for dioxin-like compounds, the

TDI is reached by people with high seafood consumption. Additionally, the results of the

third contamination scenario analysis confirmed that heavy seafood consumers can exceed

the TDI for dioxin-like compounds on the basis of seafood consumption only, even when all

seafood complies with the EU limits. This result has also been found in other studies (Baars

et al, 2004; Sirot et al, 2006).

The assessed contaminant intakes of the intake assessment based on real seafood

consumption data can also be compared with the results of other calculations:

- A French study, executed in 2006 and focussing on French high seafood consumers gave

the following results: an average iPCB intake of 57.14 ng/kg bw/day, an average intake

of dl PCB, PCDD/Fs, and total TEQ intake of 2.04, 0.62, and 2.67 pg WHO-TEQ/kg

bw/day, respectively (Sirot et al, 2006). These results are of the same order of magnitude

as the assessed intakes for the higher percentiles of the Belgian adult population.

- The British Scientific Advisory Committee on Nutrition (SACN) and Committee on

Toxicity (COT) in the UK calculated a total TEQ intake via seafood of 0.6 and 3.9 pg

WHO-TEQ/kg bw/day for an average and a heavy seafood UK adult consumer,

respectively (SACN/COT, 2004).

- A comparison can also be made with data of a recent Spanish study focussing on

contaminant intake via seafood consumption only (Bocio et al, 2007). The average

contaminant intake assessed for an adult man (70 kg) amounts to 0.46 pg WHO-TEQ/kg

bw/day for dl PCBs, 0.086 pg WHO-TEQ/kg bw/day for PCDD/Fs, and 0.54 pg WHO-

TEQ/kg bw/day for total TEQ (Bocio et al, 2007). The intake of total TEQ is quite similar

to the one assessed in our study when excluding the Baltic herring and salmon (Table

IV.2). In contrast, the assessed intake of PCDD/Fs in our study is high compared to the

Spanish study. This can partly be caused by a different seafood consumption pattern and

a different level of contamination.


155

4.1.2. Results of the intake assessment based on seafood consumption

scenarios

The results of the consumption scenario analyses show that the Belgian DRI for EPA plus

DHA can be reached through regular consumption of seafood, more specifically:

1. a combination of lean and fatty fish species (on average 50% of each) minimum three

times a week; or

2. fatty fish species two times a week.

A consumption of three times a week fatty fish, however, leads to an intake of dioxin-like

compounds close to the TDI, which is of potential toxicological concern since also other food

items, mainly of animal origin, contribute to the daily intake of dioxin-like compounds.

Again, MeHg contamination does not seem to be an issue of toxicological concern, even in

scenarios with elevated fish consumption frequencies. Hence, the consumption limits for

fish determined in this PhD-study are driven by the presence of dioxin-like contaminants,

which was also concluded by Foran et al (2005) performing an analysis of the risks and

benefits related to salmon consumption.

Despite the fact that it is possible to meet the EPA plus DHA recommendation by consuming

twice a week fatty fish without exceeding the TDI for dioxin-like compounds, many

obstacles at the consumer level exist to convince people to consume seafood twice a week.

Low perceived convenience, high price perception, and low liking of fish taste by one of the

family members act as major barriers to increase seafood consumption in Belgium (Olsen et

al, 2007). In chapter V of this PhD-thesis, some more attention is given to the perception of

Belgian consumers concerning seafood.

Due to the existing barriers to increase seafood consumption, it was worth investigating

what can be the role of EPA plus DHA enriched food items as a dietary source of LC n-3

PUFAs. The results showed that regular seafood consumption (twice a week), including fatty

fish species, in combination with regular consumption of EPA plus DHA enriched

margarine can be advised to safely increase the LC n-3 PUFA intake. Apart from margarines,

LC n-3 enriched eggs are supplied by two different brands on the Belgian market. The first

brand stated that an enriched egg contains 110 mg of EPA plus DHA. The second reported a

concentration of 125 mg DHA per egg. The mean weight of a normal egg is assumed to be 60

g (Belgian Health Council, 2005). On the basis of the most recent Belgian Food Consumption


156

Survey (De Vriese et al, 2006), it is known that the Belgian adults consume on average 10.0 g

egg/day, i.e. one egg a week, with the 97.5th percentile equal to 31.9 g/day, i.e. 3 to 4 eggs a

week (De Vriese et al, 2006). Assuming that consumers would all eat EPA and DHA enriched

eggs (110 mg EPA plus DHA/egg), this would lead to an average daily intake of 18.3 mg

EPA plus DHA, being 2.7 % of the DRI. To reach the DRI of 681 mg EPA plus DHA a day,

consumers should eat six eggs a day, increasing the cholesterol intake to 1483.2 mg/day (412

mg cholesterol/egg), whereas the Belgian recommendation states to reduce the cholesterol

intake to a maximum of 300 mg/day (Belgian Health Council, 2007). This indicates that the

contribution of EPA plus DHA enriched eggs to the total intake is low, due to the rather low

EPA and DHA concentration and the limited consumption of eggs. Enriched eggs can help to

increase the LC n-3 PUFA intake, but they can not be advised as only or major source to

achieve the EPA plus DHA DRI. Nevertheless, it must be admitted that the use of eggs in

prepared food items as cakes and pastries are not taken into consideration in this estimation,

which leads to an underestimation. Moreover, apart from margarines and eggs the supply of

omega-3 enriched food items and supplements is currently increasing. However, currently

there is still a lack of information about the EPA and DHA concentration of these food items

and supplements as well as about the current consumption rate of these products. Further

research is needed to take also this new fortified food items into account.

4.2. Risk and benefit – reference values

At this moment, no common currency exists to evaluate benefits and risks in one single step.

In other words, no common scale of measurement exists to compare human health risks and

benefits, hampering a complete and in depth risk-benefit analysis. Attempts have been made

to combine both assessments in terms of QALY’s, i.e. quality-adjusted life years (Cohen et al,

2005; Ponce et al, 2000), but many uncertainties remain to be solved in order to make a broad

application of this procedure possible. The largest uncertainties are associated with the dose-

response relationships (Cohen et al, 2005). Moreover, both published QALY-investigations

related to seafood consumption did not take into account dioxin-like contaminants, but

focused on MeHg only (Cohen et al, 2005; Ponce et al, 2000). In our approach, DRIs and TDIs

are used for the evaluation of human health benefits and human health risks, respectively,

but these values were determined taking into consideration different end points and they

refer to the intake via the total diet, while in this chapter they were used to evaluate the


157

intake via seafood consumption only. In the future, a more common methodology should be

developed to go beyond these boundary conditions. Yet, in the meantime an attempt was

made to describe the situation as accurately as possible, by means of an intake assessment of

multiple nutrients and contaminants. This answered the need recently indicated by Domingo

et al (2007a; 2007b), stating that it is of high public health interest to calculate seafood

consumption limits taking into account multiple contaminants and multiple seafood species

on the level of overall diets.

It is clear that the executed evaluation heavily depends on the TDI and DRI values applied.

Differences in the actual choice of the TDI and DRI will lead to differences in seafood

consumption recommendations independent from differences in the overall data used. For

MeHg, we applied a TDI of 0.229 µg/kg bw/day (TWI of 1.6µg/kg bw/day) as proposed by

the Joint FAO/WHO Expert Committee on Food Additives (JECFA) taking into account the

latest epidemiological results with regard to developmental toxicity (EFSA, 2004; World

Health Organization, 2007). Also the SACN/COT adapted the JECFA reference value of 1.6

µg/kg bw/week but only when assessing the dietary exposure of pregnant women and

women who may become pregnant within the following year (SACN/COT, 2004). Within

the context of a risk-benefit analysis concerning seafood consumption, SACN/COT

proposed to apply an intake limit of 3.3 µg MeHg/kg bw/week to the rest of the population

as a TWI for the rest of the population to protect against non-development adverse effects

(SACN/COT, 2004). Since MeHg intake via seafood did not seem to be of toxicological

concern when using the most stringent limit, it will of course be evaluated in the same way

when using a higher intake limit.

For dioxin-like compounds, a TDI of 2 pg WHO-TEQ/kg bw/day is applied, as proposed by

the EU (Scientific Committee on Food, 2001). This TDI is mainly based on developmental

toxicity, e.g. effects on the developing male reproductive system resulting from maternal

exposure to dioxin-like compounds and is also considered adequate to protect against other

possible effects of dioxin-like compounds, such as cancer (non-genotoxic mechanism) and

cardiovascular effects (Scientific Committee on Food, 2001). SACN/COT also proposed a

TDI of 2 pg WHO-TEQ/kg bw/day to protect against developmental toxicity but proposed a

different guideline of 8 pg WHO-TEQ/kg bw/day (SACN/COT, 2004), which should be

appropriate when considered in relation to the most sensitive and relevant non-development

effects of dioxin-like compounds (increased cancer risk). According to SACN/COT, this


158

guideline level should be used for older women and males when considering risk-benefit

aspects of seafood consumption. The reason for this is that developmental toxicity is of less

concern in the older population, whereas the positive effects of seafood consumption on

cardiovascular disease become more important. The application of the latter TDI would lead

to a lower percentage of people being at risk to exceed the TDI and to a higher consumption

limit for fatty fish.

A similar discussion is relevant for the DRI used to evaluate the EPA plus DHA intake. The

EPA plus DHA recommendation in Belgium seems to be quite high when compared to other

countries, i.e. 0.3 % of the total energy intake per day (Belgian Health Council, 2007). For the

adolescent and adult population considered in this intake assessment study, this yields a

reference value for EPA plus DHA of 720.5 mg/day and 682 mg/day, respectively. In

comparison:

- in the Netherlands, 200 mg EPA plus DHA per day is recommended (Kromhout, 2001);

- in France, the DRI for EPA and DHA is 0.2 % of the total energy intake, with a minimum

of 0.05 % contributed by DHA (Legrand et al, 2001), estimated to be equal to 500 mg/day

for French men and 400 mg/day for French women;

- in Germany, a daily intake of 350 mg LC n-3 PUFA is recommended (Bauch et al, 2006);

- in the UK, SACN/COT recommends a LC n-3 PUFA intake of minimal 450 mg/d

(SACN/COT, 2004);

- in the United States, the American Heart Association (AHA) formulated a dietary

recommendation of 500 mg/day of EPA plus DHA for cardiovascular disease risk

reduction. For patients with documented coronary heart disease, the AHA recommends

1 g of EPA plus DHA per day (Gebauer et al, 2006; Kris-Etherton et al, 2003).

- the International Society for the Study of Fatty Acids and Lipids (ISSFAL) recommends

to healthy adults a minimum intake of 500 mg per day for EPA and DHA for

cardiovascular health (ISSFAL, 2007).

Application of a lower reference value to evaluate the EPA plus DHA intake as assessed in

this PhD-study would increase the percentage of individuals having an intake above this

level and would lead to the conclusion that (1) a consumption of seafood twice a week,

varying between lean and fatty species (Fig. IV.4 and IV.5), and (2) combination of once a

week seafood with regular use of LC n-3 enriched margarine (Fig. IV.6) would be sufficient

to reach the EPA plus DHA intake recommendation.


159

5. Conclusions

The probabilistic intake assessments executed with the ProbIntakeUG-module gives better

insight in the problematic and complex nature of seafood consumption including health

benefits as well as risks. The results make clear that the intake of fat-soluble nutrients and

contaminants is highly correlated, which is an important fact that has to be taken into

consideration when formulating public food and health recommendations.

The current recommendation of the Belgian Health Council is to consume one or two

portions of seafood per week, corresponding to 150 to 300 g of seafood per week (Belgian

Health Council, 2004b). The AHA recommends that adults should eat fish (particularly fatty

fish) at least two times a week (Kris-Etherton et al, 2003).

On the basis of both consumption databases (adolescents and adults), the simulation results

predicted that both populations currently do not reach an adequate intake for the three

nutrients considered, at least when only seafood consumption is accounted for. This is

mainly due to low frequency of seafood consumption. Regarding the contaminants, (Me)Hg

contamination of seafood assumed to be available on the Belgian market is not a major issue.

In contrast, exceeding the TDI for dioxin-like compounds was noticed for heavy seafood

consumers.

Combination of regular seafood consumption (twice a week), with important contribution

of fatty fish species (at least 50%), in combination with regular consumption of EPA plus

DHA enriched margarine can be advised to maximize LC n-3 PUFA intake without

exceeding the TDI for dioxin-like compounds. Meanwhile, vitamin D and iodine intake

will increase as well. It is important to add that no other dietary sources of dioxin-like

compounds were taken into account in this assessment. Some information about the

contribution of seafood to the total dietary intake of dioxin-like compounds is given in

chapter V, part 1.

It should, however, be kept in mind that the conclusion “seafood consumption twice a week

would not lead to contaminant intakes of major toxicological concern” is essentially

conditional upon compliance of existing structural and rather strict rules and regulations,

and that extensive control programs aiming to prevent that highly contaminated food items


160

(above the EU limits) become available for consumers. In other words, the positive

conclusion from this study should in no way be interpreted as an argument in favour of any

weakening or downplaying of current regulations and monitoring programmes.

Chapter V. General discussion

161


This last chapter starts with summarizing the main findings of this PhD-thesis. Briefly, the

current intake of long chain omega-3 fatty acids (LC n-3 PUFAs) by the Belgian population is

too low compared to the recommendations and compared to the levels necessary to lead to

cardiovascular benefits (Chapter II). The most direct approach to correct this deficiency is to

increase consumption of foods rich in LC n-3 PUFAs, mainly seafood (Harris, 2007). Indeed,

the American Heart Association recommends adults to consume at least two preferably fatty

fish meals per week (Kris-Etherton et al, 2003). Similar recommendations have been made by

health authorities in other countries, e.g. the United Kingdom (SACN/COT, 2004).

Before formulating similar recommendations for the Belgian population, it was investigated

whether the LC n-3 PUFA recommendations could be reached without increasing the intake

of contaminants via seafood to levels of toxicological concern. It was concluded in chapter IV

that the Belgian population can consume fatty fish up to twice a week without exceeding

tolerable daily intake (TDI) levels for mercury and dioxin-like compounds. However, this

evaluation did not consider intake of dioxin-like compounds from other dietary sources, i.e.

mainly other foods from animal origin. In this chapter, the importance of other food items to

the total intake of dioxin-like compounds is briefly described.

Considering the current low consumption of seafood by a large part of the Belgian

population, major dietary shifts will be needed in the future to achieve increased seafood

consumption. Additionally, different barriers exist related to increased seafood consumption

and these can not be neglected in this PhD-thesis. Two main barriers are discussed:

consumer perception about seafood and sustainability issues. Related to the sustainability

question, aquaculture and non-seafood sources of LC n-3 PUFAs are briefly introduced and

discussed. Next, some aspects, opportunities, and disadvantages of omega-3 supplements

and fortified foods are discussed. The chapter concludes with some thoughts for future

research.


162

1. Main findings

Chapter II of this PhD-thesis presented the results of an intake assessment of individual

PUFAs via the total diet for three different subgroups of the Flemish population and an

intake assessment of vitamin D via the total diet for Flemish adolescents. The following

PUFAs were considered: linoleic acid (LA), α-linolenic acid (LNA), arachidonic acid (AA),

eicosapentaenoic acid (EPA), docosapentaenoic acid (DPA), and docosahexaenoic acid

(DHA). First, the results made clear that a lower intake of LA and a higher intake of LNA

are necessary to decrease the LA/LNA ratio, since a lower LA/LNA ratio can help in the

prevention of some chronic diseases. This can be reached by a higher consumption of LNA

rich foods e.g. by replacement of omega-6 (n-6) rich oils by omega-3 (n-3) rich oils such as

linseed and rapeseed in food formulations. Second, dietary shifts are necessary to bridge the

gap between the intakes and the recommendations of LC n-6 and n-3 PUFAs. Regular

replacement of meat products rich in saturated fatty acids (SFAs) by poultry meat is a

possible solution to increase the AA intake, which was evaluated to be very low for Flemish

pre-school children. Moreover, this study suggests that seafood and particularly fatty fish

consumption should be stimulated in all subgroups of the population since it is a rich

source of LC n-3 PUFAs of which the current intakes were evaluated to be far below the

recommendations. Moreover, increased seafood consumption can lead to higher vitamin D

intake and can replace SFA-rich food items. In spite of the fact that iodine was considered as

a nutrient in the probabilistic intake assessment of nutrients and contaminants via seafood

consumption only, no assessment of iodine intake via the total diet was performed, as a lot of

the available food consumption databases lack good information about the iodine

concentration in food items.

However, when recommending higher seafood consumption, one may not neglect the

nutritional-toxicological conflict related to seafood consumption. This conflict arises from

the fact that increased consumption of seafood will at the same time increase the intake of

environmental contaminants like methyl mercury (MeHg), PCBs, and dioxin-like (dl)

compounds (dl PCBs and PCDD/Fs). Therefore, a methodology was developed in this PhD-

thesis, including the elaboration of databases and the development of a software module in

order to execute a probabilistic assessment of the simultaneous intake of multiple

compounds (LC n-3 PUFAs, vitamin D, iodine, (Me)Hg, PCBs, and dioxin-like substances)

via seafood consumption.


163

Subsequently, the assessed intakes were evaluated to determine:

1. if nutrient intakes reached the recommendations, and

2. if contaminant intakes did not exceed a level of toxicological concern.

Two different situations were studied:

1. Starting form the current seafood consumption data of two different subgroups of the Belgian

population (adolescents and adults).

Based on these consumption data, the simulation results predicted that the studied

populations did not reach a sufficiently high intake for the three nutrients under

consideration (LC n-3 PUFAs, vitamin D, iodine) taking into account only seafood

consumption. Regarding the contaminants, MeHg contamination of seafood on the Belgian

market did not seem to be an issue of toxicological concern. In contrast, the TDI for dioxin-

like compounds was exceeded in the case of high seafood consumers on the basis of their

seafood consumption only.

2. It was investigated whether the Belgian recommendation for LC n-3 PUFAs (i.e. 0.3 % of total

energy intake) could be reached through seafood consumption, without exceeding TDIs of MeHg

and dioxin-like compounds. Also the contribution of LC n-3 enriched margarines was assessed.

The results indicated that the Belgian recommendation for EPA plus DHA can be reached by

consuming fatty fish twice a week, or by varying between lean and fatty fish minimally

three times a week. It was found that MeHg intake was not an issue of toxicological concern

for the Belgian population at this level of seafood consumption. The intake of dioxin-like

compounds from seafood approximated the TDI when consuming fatty fish three times a

week or more, being a potential toxicological risk since also other food items contribute to

the daily intake of dioxin-like compounds. Use of LC n-3 enriched margarine can further

help to increase the LC n-3 PUFA intake, on average by 159 mg/day. To conclude,

combination of regular seafood consumption (twice a week), with important contribution of

fatty fish species, in combination with regular consumption of LC n-3 enriched margarine

can be advised to maximize LC n-3 intake. This quite positive message related to seafood

consumption is in line with the conclusion of a recent review on the risks and benefits of

seafood consumption (Mozaffarian & Rimm, 2006). The conclusion of this review was that

for major health outcomes among adults, the benefits of seafood consumption exceed the

potential risks. They made a separate conclusion for women of childbearing age, namely that


164

also for these group benefits of modest seafood consumption, excepting a few selected

species, also outweigh risks (Mozaffarian & Rimm, 2006).

There is, however, an important limitation related to this study, namely that it was focused

on seafood. As a result contaminant intake from other food sources was not taken into

account in this PhD-thesis. For MeHg this is not a problem since exposure to MeHg occurs

almost exclusively through consumption of seafood (Clarkson & Magos, 2006). In contrast,

consumption of other food items, mainly from animal origin, contributes to the intake of

dioxin-like compounds. Recent research executed at the Department of Public Health (Ghent

University) assessed the dietary exposure to dioxin-like compounds via the total diet in

three age groups on the basis of data from the Flemish Environment and Health study. The

intake of dioxin-like compounds via animal fat of various sources was assessed. In total, 1636

adolescents (14-15 years), 1186 mothers (18-44 years), and 1586 adults (50-65 years)

participated in the study and completed a semi-quantitative food frequency questionnaire

(FFQ). Individual consumption data were combined, via a simple distribution approach,

with recent data on PCDD/Fs and dl PCBs in food items, available on the Flemish market. It

was found that seafood was the most important contributor, counting for 25.0, 29.4, and

43.3% in the group adolescents, mothers, and adults, respectively (Bilau et al, 2007). It is

important to note that these percentages are averages per age group. On individual level,

there is a high influence of the overall dietary pattern. For example, people regularly

replacing meat by seafood will have a higher contribution of seafood, whereas heavy meat

consumers who never or seldom eat seafood will have a higher contribution of meat and a

low contribution of seafood. The other main contributors were in order of importance added

fats, dairy products, and meat and meat products (Bilau et al, 2007).

2. Consumer issues

When aiming to increase seafood consumption in a population, it is important that the

consumers are convinced of the importance to do so. I participated in different research

studies of the Department of Agricultural Economics (Ghent University), investigating the

perception of Belgian consumers about benefits and risks of seafood, farmed and wild fish,

and sustainable and ethical issues. The following papers describe the results of these studies:


165

1. Verbeke W, Sioen I, Pieniak Z, Van Camp J, De Henauw S. Consumer perception versus

scientific evidence about health benefits and safety risks from fish consumption. Public

Health Nutrition 2005; 8(4): 422-429.

2. Verbeke W, Sioen I, Brunsø K, De Henauw S, Van Camp J. Consumer perception versus

scientific evidence of farmed and wild fish: exploratory insights from Belgium.

Aquaculture International 2007; 15(2): 121-136.

3. Verbeke W, Vanhonacker F, Sioen I, Van Camp J, De Henauw S. Perceived importance of

fish ethics and sustainability: a consumer behavior perspective. Ambio 2007; In Press.

The main findings of these studies being important in the light of this PhD-thesis are

summarized below.

2.1. Consumer perception about benefits and risks from seafood

consumption

Several studies indicate that fish and other seafood is strongly perceived as a healthy food by

consumers, particularly as compared with meat as its main substitute protein source

(Brunsø, 2003; Gross, 2003). However, one of the potential barriers to eat fish more

frequently may pertain to safety risks (Leek et al, 2000; Trondsen et al, 2003; Verbeke &

Vackier, 2005).

This first mentioned Belgian study of Verbeke et al (2005) aimed

1. to investigate consumer attitude and perception towards fish, and

2. to explore the potential gap between scientific evidence versus consumer perception

related to fish consumption benefits and risks.

A self-administered questionnaire was used and the study sample consisted of 429 adults

being the main person responsible for food purchasing in the household (284 women and

145 men, aged 18-83 years).

The results of this study showed that consumers believe with reason that fish is healthy and

nutritious, being an interesting result when aiming to convince consumers to eat regularly

seafood. Nevertheless, 43% of the respondents of this study reported not to eat fish at least

once a week. The consumers’ knowledge of the vitamin D content in fish was good, but the

results with respect to n-3 PUFAs showed that most people are not fully aware of the specific


166

nutrient content of seafood. The latter was a quite surprising result and showed that there is

a definite need to inform people about the nutritional value and benefit of seafood. A

majority of consumers scored neutral on the belief that fish is safe, probably being a result of

the messages in the media about contamination of seafood. Attitude towards fish was found

to include positive versus negative perceived attributes. In general, the taste and the healthy

image of fish were two well-appreciated characteristics which can be used when promoting

fish consumption. However, the bones in fish and the price are identified as the most likely

attitudinal barriers to more frequent fish consumption. In general, the healthy image of fish

prevails over its image of being potentially unsafe. This study exemplified the need for

nutritional education and more effective communication about seafood to the broader

public (Verbeke et al, 2005).

2.2. Consumer perception of farmed and wild fish

Data resulting from the questionnaire mentioned in 2.1. of this chapter were also used by

Verbeke et al (2007b) to explore the perception of Belgian consumers regarding farmed

versus wild fish. In addition, three consumer focus groups were conducted with six to eight

participants in each session with the aim to gain further insights into consumer perception of

farmed versus wild fish.

The majority of the consumer sample reported no perceived differences between farmed and

wild fish. However, mean perception scores were slightly in favour of wild fish on the

attributes taste, health, and nutritious value, in particular amongst consumers aged 55 years

and older. The availability of farmed fish was perceived to be better than that of wild fish,

while the consumers’ perception of safety did not differ between farmed and wild fish. The

focus group discussions indicated that consumers’ opinions and beliefs about farmed fish are

mainly based on emotion and image transfer from intensive terrestrial livestock production

rather than on awareness and factual knowledge of aquaculture (Verbeke et al, 2007b). These

results indicate that a good indication about the production method of the seafood product

as well as some information about the advantages and disadvantages of the production

method would help consumers to get a good opinion about farmed seafood. This can be

crucial when promoting farmed seafood in the future, since the contribution of aquaculture


167

to the total food supply is increasing year by year. A lack of good information can result in a

wrong opinion of consumers being negative for their consumption attitude.

2.3. Consumer perception about ethical and sustainability issues of fish

Although sustainability and ethics are of increasing public importance, little research has

been conducted to reveal its association with fish consumer behaviour. In a third study,

Verbeke et al (2007a) collected data through a postal self-administered survey from a sample

of 381 Flemish women aged 20 to 50 years, with the objective to verify if ethical matters

contribute to choosing or rejecting either farmed or wild fish. In addition, characterisation

was made of persons who show more interest in ethical matters and sustainability, based on

socio-demographics, behaviour, and attitude towards fish and involvement in fish

consumption.

The results indicated that ethical and sustainability issues were indicated as being important

by the consumer in relation to fish consumption. This claimed importance, however, was

neither translated in a significant correlation with total fish consumption frequency nor with

general attitude towards eating fish. The choice not to eat wild fish seemed to have part of its

origin in ethical and sustainability issues, given a significantly higher importance attached to

these issues by consumers who claim to refuse wild fish, and a negative correlation between

interest in sustainability and cod consumption. This was not the case with respect to rejecting

farmed fish, which was not associated with importance attached to ethical or sustainability

issues. Finally, respondents with the highest importance attached to ethical and

sustainability issues are the ones with the highest interest in information in general, with the

highest belief in a potential benefit of receiving more information, and the highest perceived

efficacy of their own deliberate choice (Verbeke et al, 2007a).


168

2.4. Conclusion related to consumer perception about seafood

The above summarized studies performed amongst Belgian consumers indicated that a gap

exists between consumers’ knowledge and scientific facts about seafood. This is mainly

caused by a lack of information. Moreover, confusion among consumers is probably

explainable by the conflicting messages about seafood in the media last years. Efforts to

inform consumers about the composition and the origin of seafood products in an

understandable and univocal way are therefore crucial when aiming to convince them to eat

seafood on a more regularly basis. In contrast, other consumer barriers as the presence of

bones and the price of seafood are more difficult to eliminate.

3. Sustainability

3.1. Depletion of natural fish stocks

There are concerns that the increased consumption of seafood and the increased use of fish

oils for enriched foods are not a sustainable solution from an ecological point of view. The

worldwide increase in consumption of seafood and derived seafood products during recent

decades are mainly due to recommendations that seafood is part of a healthy human diet

(World Health Organization, 2003), the increasing world population, higher living standards,

and the overall image of seafood among consumers (Brunsø, 2003; Cahu et al, 2004; Verbeke

et al, 2005). In 2004, total capture fisheries and aquaculture supplied the world with 106

million tonnes of seafood for human consumption, providing an apparent per capita of 16.6

kg (FAO, 2007). The increase in demand and supply has led to an expansion of the fishing

fleet. Together with higher fish capture efficiency, this has contributed to overfishing and the

risk of depletion of some natural fish stocks. As such, seafood species like Alaska Pollack,

blue whiting, tuna, and mackerel became globally limited food sources (Bauch et al, 2006;

European Commission, 2007). Moreover, the FAO (2007) reported that total catches in 2004

decreased by over 10% in comparison with 2002 in the Northeast Atlantic area, which was

estimated to be the most important natural source for seafood on the Belgian market

(Chapter III).


169

3.2. Aquaculture: a valuable alternative?

As an alternative to wild caught seafood, consumers are offered now farmed seafood. In

2004, aquaculture accounted for 43% of the total world seafood supply and the contribution

of aquaculture will continue to expand, while marine capture fisheries seem to have reached

a ceiling (FAO, 2007). It is predicted that in 2030 aquaculture will provide half of the total

amount of seafood consumed worldwide (European Commission, 2007; Tidwell & Allan,

2001). Therefore, it is a challenge to develop sustainable aquaculture, which in the same time

minimizes bioaccumulation of contaminants and other food safety risks, as the risk of

contamination by chemical and biological agents is greater in freshwater and coastal

ecosystems than in open seas (McMichael & Butler, 2005; World Health Organization, 1999).

3.2.1. Wild versus farmed seafood

To set both ways of seafood supply side by side, a limited comparison of wild and farmed

seafood is given, considering nutritional as well as food safety aspects. The nutrient content

may differ between wild and farmed seafood species, viewed the difference in their diet.

Wild seafood eats plankton, small algae, small fishes, and other seafood species. Their diet is

affected by environmental and seasonal changes. This will affect the proximate composition

of the muscle. In aquaculture, farmed seafood is provided with a supply of nutrient-dense

formulated feed, which is constant throughout the year and which enables them to deposit

large reserves of lipids. Moreover, the content of the formulated feed can be changed in

function of the wishes of the producer and the consumer. Different studies indicated that the

lipid content of farmed fish is generally larger than that of their free-living counterparts and

the levels of EPA and DHA are generally lower, when expressed relatively to total fatty acids

(Bandarra et al, 1997; Haard, 1992; Nettleton & Exler, 1992; Olsson et al, 2003; Sérot et al,

1998). But viewed their higher lipid content, the amounts of EPA and DHA provided by a

given quantity of farmed fish may be higher than in the same amount of wild fish. On the

other hand, cholesterol and protein levels are similar in farmed and in wild fish. Nettleton &

Exler (1992) showed that levels of different vitamins (vitamin A, C, B12, B2, B3, B5, B8, and

folic acid) were similar or higher in farmed species. Cahu et al (2004) concluded that the

nutritional content of farmed fish is at least as beneficial as that of wild fish and farmed fish


170

also has advantages of freshness because the storage conditions between slaughtering and

sale are more verifiable.

Next, food safety problems are linked to aquaculture. Regarding the presence of

environmental contaminants, the most important contaminant source for farmed seafood is

their formulated feed (Bell et al, 2005; Karl et al, 2003). Some recent studies showed that

concentrations of organochlorine contaminants are significantly higher in farmed salmon

than in wild and that farmed salmon of Europe are significantly more contaminated than

farmed salmon from South and North America (Easton et al, 2002; Hites et al, 2004a). In

contrast, the European Food Safety Authority (EFSA) conducted a scientific assessment of

the health risks related to human consumption of wild and farmed fish and concluded that

with respect to their safety for the consumer there is no difference between wild and farmed

fish (EFSA, 2005b). Nevertheless, aquaculture industries should be able to take steps to

reduce contaminants in fish feed by applying purification processes and as such be able to

provide the population with naturally rich dietary LC n-3 sources, being safe from a

toxicological point of view. Another food safety issue related to aquaculture is the use of

additives like colorants and the use of drugs (Alderman & Hastings, 1998). However, due to

the large expansion of global aquaculture, countries worldwide have implemented a large

number of aquaculture regulations to control inadequate developments for example related

to drug use (FAO, 2007).

Nevertheless, these current aquaculture regulations cannot guarantee sustainability,

especially as most of them focus on the individual farmer and do not consider the additive or

synergetic effect of multiple farms on a particular area. At the same time, farmers’ economic

appraisals tend to have a narrow view, which do not include the medium- and long-term

revenues and costs that may be imposed on the farming activity itself and on the rest of the

society in the form of a reduced supply of ecosystem goods and services (FAO, 2007).

3.2.2. Feeding farmed seafood

A final aspect related to aquaculture is the widespread practice of feeding wild-caught

seafood to farmed piscivorous seafood species, e.g. salmon. This practice decreases the

availability of seafood for direct human consumption on a per capita basis (McMichael &


171

Butler, 2005; Naylor et al, 2000). This leads to a paradox: aquaculture is a possible solution,

but also a contributing factor to the collapse of fisheries stocks worldwide (Naylor et al,

2000). It must be said that up to now, aquaculture is still the largest consumer of fish derived

oils, and so is clearly not capable of operating in a sustainable manner (Napier & Sayanova,

2005). So, to meet consumer demand for piscivorous fish and at the same time reduce the

pressure on wild fish stocks, land-based plants - such as grains, soybeans, legumes, and

blended vegetable oils - are increasingly being used as staple fish food. The disadvantage is

that using this feed will lower the LC n-3 PUFA concentration compared to the wild-caught

species, affecting one of the important nutritional advantages of seafood consumption

(McMichael & Butler, 2005). However, actually different studies are running that investigate

whether a combination of a diet high in vegetable oil with a diet high in fish oil only during

the last weeks of the production cycle can lead to a restoration of EPA and DHA

concentration to equal levels when compared to fish oil fed counterparts (Bell et al, 2005;

Napier & Sayanova, 2005). More research is needed and probably it will take time before

such customs are used in practice. An alternative to this kind of combined diets can be the

use of single cell oils and genetically modified plants (and animals) as sustainable sources of

LC n-3 PUFAs. An introduction to these rather new developments is given in the paragraph

below.

4. Potential sustainable sources of LC n-3 PUFAs

4.1. Single cell oils

As a more sustainable source of LC n-3 PUFAs compared to fish and fish oils, commercial

companies started with the production of microbial oils, otherwise referred to as single cell

oils (SCO), based on a DHA producing organism that can grow in large-scale industrial

fermentors. Today, there are at least three different fermentation processes commercially

available, each using a different microorganism for the production of DHA. For example,

Crypthecodinium cohnii produces a triacylglycerol oil in which DHA can reach levels between

40 and 50% of the total fatty acids and, moreover, occurs as the sole PUFA. Currently, the

major application of these SCOs is in the form of a supplement for infant nutrition (Ratledge,

2004).


172

4.2. Metabolic engineering in plants (and animals)

A second alternative way to increase the supply of LC n-3 PUFAs is through plant

technology, in other words metabolic engineering in plants which manipulates the lipid

metabolism and leads to the production of novel plant oils (Kinney, 2006; Voelker & Kinney,

2001). Advances in gene expression and genomics have recently led to the engineering of

multigene pathways in plants, such as microbial LC PUFAs biosynthetic pathways

reconstituted in the seeds of crop plants (Domergue et al, 2005). In other words, identification

of a gene cluster coding the entire pathways of PUFA biosynthesis in the microorganisms

used in the SCO applications opened the way to clone this contiguous gene sequence into

various plants in order to achieve PUFA production in agronomically important crops such

as sunflower, soybean, or rapeseed (Ratledge, 2004). Application of this technology on a

larger scale would provide a sustainable source of LC n-3 PUFA for use in human food and

animal feed (for terrestrial as well as aquatic animals). A lot of research is done.

Nevertheless, more efficient techniques are still required to optimize the LC n-3 PUFA

synthesis in plants (Domergue et al, 2005; Napier & Sayanova, 2005). It is also worth to

mention that very recent studies report on the development of transgenic pigs able to

convert n-6 PUFAs to n-3 PUFAs (Prather, 2006). However, a long research and regulatory

way will be needed before they will enter the food chain, if ever they will enter, because lot

of controversy still exists about the presence of transgenic foods on the market. Discussion of

the latter is beyond the scope of this PhD-thesis.

5. Omega-3 supplements and omega-3 fortification

Apart from seafood, fish oil supplements and food items fortified with LC n-3 PUFA are

currently available as alternative LC n-3 PUFA dietary source. Supplements are usually

taken additionally to the normal diet. In contrast, fortified foods will normally replace

certain unfortified food items in the normal diet. Fortification of food items can be done in

different ways: direct and indirect fortification. A short discussion with respect to these

alternative LC n-3 sources is given in this paragraph.


173

5.1. Fish oil supplements

In addition to being used in the food and animal feed industry, fish oils have also been used

traditionally as dietary supplements. The dietary supplement market expanded significantly

and includes these oils in a variety of different formulations (Jacobs et al, 2004). Most known

are the encapsulated fish oils sold as dietary supplements for their high content of EPA and

DHA. The supply of these supplements on the Belgian market is increasing and, thus,

consumption of these supplements is increasing as well, mainly due to good promotion

campaigns and the ‘commercial hype’ that is currently related to omega-3 fatty acids. Fish

oils are derived from fish liver (e.g. cod liver oil) or from muscle tissue of fatty fish species.

However, fish oils represent a large accumulation potential for lipophilic persistent organic

pollutants present in the marine ecosystem (Jacobs et al, 2004). Therefore, it is important to

monitor closely the levels of organic pollutants that might be present (Covaci et al, 2006;

Fernandes et al, 2006; Jacobs et al, 2004; Storelli et al, 2004). In an effort to remove these

contaminants, the industry employs a number of refining processes. Molecular distillation is

very effective in removing halogenated contaminants as PCBs, but it removes at the same

time beneficial components like EPA and DHA. Thus, some manufactures have invested in

newer purification methods that are able to achieve this balance and are claiming success

(Fernandes et al, 2006).

Several recent papers are available reporting the contamination of fish oil dietary

supplements available on the European market (Covaci et al, 2006; Fernandes et al, 2006;

Jacobs et al, 2004; Storelli et al, 2004). Storelli et al (2004) analysed dietary supplements based

on cod-liver oil from the Italian market on PCBs, hexachlorobenzene, and chlorinated

pesticides. Jacobs et al (2004) reported analytical results of selected contaminants, including

PCBs, organochlorine pesticides, and polybrominated diphenyl ethers (PBDEs) for a range of

commercially available cod liver oil and other fish oil dietary supplements on the market in

London, UK, sampled in December 2001 and January 2002. Fernandes et al (2006) analysed

dioxins and PCBs in fish oil supplements sourced from retail outlets in the UK sampled

during 2001-2002. Covaci et al (2006) analysed fish oil dietary supplements collected in 2006

and reported the levels of PCBs and (methoxy)polybrominated diphenyl ethers. They choose

the samples as such that they covered the products available for sale in the Belgian market;

and, additionally, fish oil dietary supplements bought in the Netherlands, Ireland, United

Kingdom, and South-Africa were also analyzed. Only a summary of the concentration of


174

PCBs, dl PCBs, and PCDD/Fs are given here, since the other contaminants reported in the

articles were not considered in this PhD-thesis.

In the Italian cod-liver oil supplements, PCB concentrations ranging from 25 to 201 ng/g

lipid were found (Storelli et al, 2004). Jacobs et al (2004) reported a concentration range for

the sum of the seven indicator PCBs (iPCBs) equal to 86.7 to 133.3 ng/g oil and not detected

(ND) to 49.0 ng/g oil for cod liver oils (n=7) and other fish oils (n=6), respectively. Fernandes

et al (2006) having analysed a total of 33 dietary supplements, compromising mostly cod liver

oils but also other fish oils sampled during the same period and from the same market found

iPCB concentrations ranging from 8.3 to 266.6 ng/g oil, so being more extreme than what

Jacobs et al (2004) found. Also Covaci et al (2006) sampled 12 fish oil dietary supplements

from the UK market, but later in time, i.e. 2006. They reported a PCB concentration (sum of

20 congeners) range for these samples from <0.3 to 22 ng/g oil, being much lower than the

older samples reporting only the sum of seven iPCBs. This can confirm the downward trend,

previously reported (Fernandes et al, 2006; Jacobs et al, 2004). Further reasons for the low

contaminant levels include improved sourcing and nutritional profile development to create

specific ‘optimum’ essential fatty acid balances (Covaci et al, 2006). Furthermore, Covaci et al

(2006) found PCB concentration ranges equal to <0.3-57, <0.3-60, <0.3-95 ng/g oil

respectively for samples on the market in Belgium (n=27), The Netherlands (n=17), and other

countries (n=13).

Only Fernandes et al (2006) reported concentration ranges for PCDD/Fs and dl PCBs equal

from 0.18 to 8.4 pg TEQ and from 1.1 to 41 ng/g oil in 33 supplements collected on the UK

market between 2001 and 2002. When manufacture-recommended doses were applied to the

observed levels, the estimated upper bound human exposure to PCDD/Fs and dl PCBs

ranged from 0.02 to 7.1 pg TEQ/kg bw/day (Fernandes et al, 2006). Currently, the European

Commission legislation has set a maximum limit of 2.0 pg TEQ/g fat for PCDD/Fs and 10.0

pg TEQ/g fat for the sum of PCDD/Fs and dl PCBs for marine oils (fish body oil, fish liver

oil and oils of other marine organisms intended for human consumption) (European

Commission, 2006). Much of the samples analysed by Fernandes et al (2006) exceed these

maximum limits. But, the decreasing trend in contamination load seen previously and better

purification processes will probably cause lower concentrations in the samples currently

available on the market.


175

However, some controversy exists related to the use of supplements. First, there are

indications that dietary patterns rather than individual nutrients might account for more

important beneficial effects, taking into account the possibility of diverse interactions

between different food components (Reinert et al, 2007). In particular, opponents of

supplements and functional foods respond that it is the total diet that is important for health,

not the so-called ‘magic bullets’ (Lawrence & Rayner, 1998). Second, using supplements can

be seen as adding something extra on top of the usual diet, whereas it would be – in a lot of

cases – more beneficial to replace less healthy food items by healthy food items instead of

adding supplements to a rather unhealthy diet or - in general - unhealthy way of life. As fish

and other seafood is a good source of high quality proteins and fatty acids, as well as of

vitamin D and micronutrients, it can serve as an excellent substitute for protein sources high

in saturated fats. So, in absence of seafood allergy or dislike, seafood consumption can be

advised rather than supplement use.

5.2. Direct fortification

Several functional foods, including margarines and dressings, to which LC n-3 PUFAs are

added, have been developed in the last years (Patch et al, 2005b; Patch et al, 2005a). These

food items belong to the group of direct fortified foods since an extra nutrient, in this case

EPA and DHA, is added to the product at the end of the production process. This became

possible since technology made it achievable to produce high quality, deodorized, and

stabilized oils enriched in LC n-3 PUFAs (Harris, 2007). Currently, processed fish oils are

used as a source of LC n-3 PUFA, generally in microencapsuled form to fortify the foods in a

tasteless way. Nevertheless, in the future SCOs and genetically modified plants will probably

supply the necessary fatty acids to enrich all kind of food products. Different research

groups already investigated the impact of foods enriched with LC n-3 PUFA on

cardiovascular risk factors (Baro et al, 2003; Harrison et al, 2004; Liu et al, 2001; Metcalf et al,

2003; Murphy et al, 2007; Visioli et al, 2000). All of them concluded that increased

consumption of LC n-3 PUFA enriched foods may be associated with reduced risk in

cardiovascular diseases. As for supplements, no recent information is available neither on

the consumption of such foods by the Belgian population nor on their contribution to the

overall intake of LC n-3 PUFAs.


176

5.3. Indirect fortification: modifying animals’ diet

In the last decades, a lot of research has been devoted to the potential of modifying the fatty

acid composition of animal food products to better meet human nutritional guidelines.

Additionally, there is also growing interest in the animal feed industry to implement

strategies fulfilling these objectives. Modifying animal’s diet, i.e. indirect fortification, offers

by far the largest opportunities (Raes et al, 2004). In this respect, research efforts to increase

the (LC) n-3 PUFA content of terrestrial animal products by fortifying the animals’ feed

could contribute to improving the overall intake of (LC) n-3 PUFAs (Sontrop & Campbell,

2006). To improve the n-3 PUFA content and the n-6/n-3 ratio, the use of vegetable oils or

whole seeds rich in LNA and grass or grass products (the latter for ruminants) is a

worthwhile strategy. This approach mainly increases the LNA content but also offers

benefits in terms of increasing the LC n-3 PUFAs, although the latter response strongly

depends on the type of product. Adding fish oil or fishmeal to the diets of farm animals like

pigs considerably increases the deposition of both EPA and DHA. However, there are major

obstacles for using fish oil or fishmeal as feed ingredients (Raes et al, 2004; Sontrop &

Campbell, 2006). The use of single cell oils (SCO) or genetically modified crops might be an

alternative. Nevertheless, the potential human health benefits of specific enrichment

strategies also remain unknown. In a preliminary simulation study using Belgian food

consumption data, the impact of alterations in the n-6/n-3 PUFA ratio of farm animal foods

on the pattern of fatty acid intakes has been estimated (De Henauw et al, 2007). Although the

outcome was largely affected by the assumptions that were made, considerable shifts in the

intake of LNA could be achieved. Nevertheless, it was difficult to obtain shifts in the intake

of LC n-3 PUFA. More data and alternative strategies need to be considered in this respect.

6. Risks related to excessive intakes of LC n-3 PUFAs,

vitamin D, and iodine

In this PhD-thesis, one of the major goals was to explore pathways to bridge the gap between

the current low intakes and the recommendation of certain nutrients, particularly LC n-3

PUFAs, vitamin D, and iodine. However, data, information, and special attention in the

media about the benefits of those nutrients can lead to an increased supply and use of


177

supplements and fortified foods in some parts of the population, with an excessive intake of

those nutrients as a result. Therefore, it is worth to describe shortly the possible risks and

safety considerations related to very high intakes of LC n-3 PUFAs, vitamin D, and iodine.

6.1. LC n-3 PUFAs

The American IOM stated that there is up to now insufficient evidence to set a tolerable

upper intake level (UL) for LC n-3 PUFAs (Institute of Medicine, 2005). However, a working

group of the French Food Safety Agency (AFSSA) considers that the maximum permitted

intake for LC n-3 PUFAs should be 2.0 g/d, a level close to the mean values observed in

populations which consume large amounts of marine products on a daily basis. Rather than

seeing this value of 2 g/d as a safety limit for LC n-3 PUFAs beyond which there would be a

risk, it should be emphasised that there is no proven nutritional benefit in recommending a

daily intake near or above it (AFSSA, 2003). Two different safety aspects have to be kept in

mind. A first safety consideration related to the intake of LC n-3 PUFAs is the question

whether it leads to a higher risk of bleeding. LC n-3 PUFAs affect thrombosis by decreasing

platelet aggregation (Kris-Etherton et al, 2002). It is due to this antithrombotic effect that the

suggestion was made that LC n-3 PUFAs could maybe increase the risk for bleeding (Carroll

& Roth, 2002). However, clinical trial evidence suggest that if such a risk exists, the risk is

very small and not of clinical significance and that high-dose fish oil LC n-3 PUFA

consumption is safe (Bays, 2007). A second safety factor is that LC n-3 PUFAs are naturally

highly unstable and susceptible to oxidation, which may increase the risk for toxicity. One of

the most common ways to reduce oxidation is to add the antioxidant vitamin E to

supplements. This can overcome the oxidative stress to the body that is observed when fish

oils without vitamin E are consumed (Bays, 2007). Consequently, LC n-3 PUFA supplements

leading to an intake achieving the recommendations and containing an anti-oxidant can

generally be considered as safe.


178

6.2. Vitamin D

Hypervitaminosis D is characterised by a considerable increased plasma concentration of

1,25-dihydroxycholecalciferol (1,25(OH)2D3), the biologically active hormone formed from

vitamin D. As 1,25(OH)2D3 increases to toxic concentrations, hypercalciuria, hypercalcemia,

and, finally, extraskeletal calcification becomes evident (Hollis & Wagner, 2004; Institute of

Medicine, 1997). Moreover, reduced renal function and a severe depression illness have been

associated with hypervitaminosis D.

Based on toxicological data, an UL equal to 50 µg vitamin D per day was derived for adults

and children between 1 and 18 years old and is supported by the American IOM and the

European Scientific Committee on Food (Institute of Medicine, 1997; Scientific Committee on

Food, 2002). For infants up to 1 year of age, the UL is set at 25 µg vitamin D per day (Institute

of Medicine, 1997). The IOM stated that for most people, vitamin D intake from food and

supplements is unlikely to exceed this UL. However, people who are at the upper end of the

ranges of intake of both sources (food and supplements) and those with high intakes of fish

or fortified foods may be at risk for vitamin D toxicity (Institute of Medicine, 1997).

Conversely, recent data suggest that the current UL is far more restricted than needed to

avoid adverse effects of vitamin D. Based on a new risk assessment, Hathcock et al (2007)

suggest that 250 µg/d can be selected as new UL. This should safely permit increased intakes

of vitamin D at higher levels than previously recognized (Hathcock et al, 2007).

6.3. Iodine

Concerning iodine, it has been reported that a wide range of iodine intakes is tolerated by

most individuals, owing to the ability of the thyroid to regulate total body intake. Over 2

mg/day for long periods should be regarded as excessive or potentially harmful to most

people, but such high intakes are unlikely to arise from natural foods. However, the IOM set

an UL for iodine at 1100 µg/day for adults (Institute of Medicine, 2001). Hyperthyroidism

(result from excess iodine exposure) is largely confined to those over 40 years of age.

Symptoms of hyperthyroidism are rapid hearth rate, trembling, lack of sleep, and loss of

weight and strength (Strain & Cashman, 2002).


179

7. Thoughts for future research

This PhD-thesis helped to find out which recommendations about seafood consumption can

be formulated for the Belgian population in order to increase the intake of LC n-3 PUFAs to

levels that can help in the prevention of chronic Western diseases without leading to

contaminant intakes at levels of toxicological concern. However, as described in this final

chapter, a lot of other aspects have to be considered when advising increased seafood

consumption, as there are consumer perception and barriers, the aspect of sustainability, and

the contribution of other food items to the intake of dioxin-like compounds.

In the future, research will need to find out which is the most sustainable and most beneficial

way to provide enough LC n-3 PUFAs for the human population, without creating new risks

for man or for the environment. These new strategies can be based on the use of direct

fortified food items, like bread, yoghurts, meat products, … fortified with LC n-3 PUFAs.

Another possibility is to investigate in the production chain of animal food items, e.g. milk,

cheese, meat of terrestrial animals, seafood. By modifying the diet of these animals during

the production process, a modified composition of the final food item can be achieved as was

described in 5.3. of this chapter. However, such strategies still need some optimisation before

they can be used on broad scale. Moreover, it is not clear how consumers will react on an

increasing supply of such products.

To conclude, it remains a continuing challenge to find a good way to guide consumers to a

healthy eating behaviour in a world where large multinational companies control the global

food market. Only then, successful and sustainable public health nutrition programs will be

raised.

References

181

References

Ackman RG. Fatty acids in Fish and Shellfish. In: Chow CK (ed), Fatty acids in food and their

health implications. New York: Marcel Dekker, 2000: 153-175.

AFSSA. The omega-3 fatty acids and the cardiovascular system: nutritional benefits and

claims. Available at http://www.afssa.fr. Paris: AFSSA, 2003.

AFSSA. Dioxines, furanes et PCB de type dioxine: Evaluation de l'exposition de la

population française (Dioxins, furans and dioxin-like PCBs: Evaluation of the exposure of

the French population). Available at http://www.afssa.fr. Paris: AFSSA, 2005.

Agren JJ, Hanninen O. Effects of cooking on the fatty-acids of 3 fresh-water fish species. Food

Chemistry 1993; 46(4): 377-382.

Al Saghir S, Thurner K, Wagner KH, Frisch G, Luf W, Razzazi-Fazeli E, Elmadfa I. Effects of

different cooking procedures on lipid quality and cholesterol oxidation of farmed salmon

fish (Salmo salar). Journal of Agricultural and Food Chemistry 2004; 52(16): 5290-5296.

Alderman DJ, Hastings TS. Antibiotic use in aquaculture: development of antibiotic

resistance - potential for consumers health risks. International Journal of Food Science and

Technology 1998; 33: 139-155.

American Heart Association, Gidding SS, Dennison BA, Birch LL, Daniels SR, Gilman MW,

Lichtenstein AH, Rattay KT, Steinberger J, Stettler N, Van Horn L. Dietary

Recommendations for Children and Adolescents: A Guide for Practitioners. Pediatrics 2006;

117(2): 544-559.

Aminullah Bhuiyan AKM, Ratnayake WMN, Ackman RG. Nutritional composition of raw

and smokes Atlantic mackerel (Scomber scomber): Oil- ad water-solubles vitamins. Journal of

Food Composition and Analysis 1993; 6: 172-184.

http://www.afssa.fr/

http://www.afssa.fr/

References

182

Appleton KM, Hayward RC, Gunnell D, Peters TJ, Rogers PJ, Kessler D, Ness AR. Effects of

n-3 long-chain polyunsaturated fatty acids on depressed mood: systematic review of

published trials. American Journal of Clinical Nutrition 2006; 84(6): 1308-1316.

Ashizuka Y, Nakagawa R, Tobiishi K, Hori T, Ida T. Determination of polybrominated

diphenyl ethers and polybrominated dibenzo-p-dioxins/dibenzofurans in marine products.

Journal of Agricultural and Food Chemistry 2005; 53(10): 3807-3813.

Astorg P, Arnault N, Czernichow S, Noisette N, Galan P, Hercberg S. Dietary intakes and

food sources of n-6 and n-3 PUFA in French adult men and women. Lipids 2004; 39(6): 527-

535.

Baars AJ, Bakker MI, Baumann RA, Boon PE, Freijer JI, Hoogenboom LAP, Hoogerbrugge R,

van Klaveren JD, Liem AKD, Traag WA, de Vries J. Dioxins, dioxin-like PCBs and non-

dioxin-like PCBs in foodstuffs: occurrence and dietary intake in The Netherlands. Toxicology

Letters 2004; 151: 51-61.

Baeyens W, Leermakers M, Papina T, Saprykin A, Brion N, Noyen J, De Gieter M, Elskens M,

Goeyens L. Bioconcentration and biomagnification of mercury and methylmercury in North

Sea and Scheldt estuary fish. Archives of Environmental Contamination and Toxicology 2003;

45(4): 498-508.

Bandarra NM, Batista I, Nunes ML, Empis JM, Christie WW. Seasonal changes in lipid

composition of sardine (Sardina pilchardus). Journal of Food Science 1997; 62(1): 40-42.

Bang HO, Dyerberg J. Lipid metabolism in Greenland Eskimos. Advances in nutritional

research 1980; 31: 1-32.

Barbarich BN, Willows ND, Wang L, Clandinin MT. Polyunsaturated fatty acids and

anthropometric indices of children in rural China. European Journal of Clinical Nutrition 2006;

60(9): 1100-1107.

References

183

Baro L, Fonolla J, Pena JL, Martinez-Ferez A, Lucena A, Jimenez J, Boza JJ, Lopez-Huertas E.

n-3 Fatty acids plus oleic acid and vitamin supplemented milk consumption reduces total

and LDL cholesterol, homocysteine and levels of endothelial adhesion molecules in healthy

humans. Clinical Nutrition 2003; 22(2): 175-182.

Bauch A, Lindtner O, Mensink GB, Niemann B. Dietary intake and sources of long-chain n-3

PUFAs in German adults. European Journal of Clinical Nutrition 2006; 60(6): 810-812.

Bayen S, Barlow P, Lee HK, Obbard JP. Effect of cooking on the loss of persistent organic

pollutants from salmon. Journal of Toxicology and Environmental Health A 2005; 68(4): 253-265.

Bays HE. Safety considerations with omega-3 fatty acid therapy. American Journal of

Cardiology 2007; 99(6A): 35C-43C.

Belgian Health Council. Aanbevelingen en beweringen betreffende omega-3 vetzuren

(Recommendations and allegations concerning omega-3 fatty acids). D/2005/7795/1 -

http://www.health.fgov.be/CSH_HGR. 2004a.

Belgian Health Council. Vis en gezondheid bij volwassenen (Fish and health of adults).

D/2004/7795/3 - http://www.health.fgov.be/CSH_HGR. 2004b.

Belgian Health Council. Maten en gewichten. Handleiding voor gestandaardiseerde

kwantificering van voedingsmiddelen in België (Measures and weights. Manual for

standardised quantification of food items in Belgium). HGR 6545-2 -

http://www.health.fgov.be/CSH_HGR. 2005.

Belgian Health Council. Voedingsaanbevelingen voor België. Herziene versie 2006

(Nutritional recommendations for Belgium. Revised version 2006).

http://www.health.fgov.be/CSH_HGR. 2007.

Bell JG, McGhee F, Dick JR, Tocher DR. Dioxin and dixon-like polychlorinated biphenyls

(PCBS) in Scottish farmed salmon (Salmo salar): effects of replacement of dietary marine fish

oil vegetable oils. Aquaculture 2005; 243(1-4): 305-314.

http://www.health.fgov.be/CSH_HGR




References

184

Bender DA. The Vitamins. In: Gibney MJ, Vorster HH, Kok FJ (eds), Introduction to human

nutrition. Oxford: Blackwell Science Ltd, 2002: 125-176.

Bilau M, Matthys C, Baeyens W, Bruckers L, De Backer G, Den Hond E, Keune H, Koppen G,

Nelen V, Schoeters G, van Larebeke N, Willems JL, De Henauw S. Dietary exposure to

dioxin-like compounds in three age groups. Results from the Flemish Environment and

Health study. Chemosphere 2007; In Press. Available online doi: 10.1016/j.chemosphere

2007.07.008.

Bocio A, Domingo JL. Daily intake of polychlorinated dibenzo-p-dioxins/polychlorinated

dibenzofurans (PCDD/PCDFs) in foodstuffs consumed in Tarragona, Spain: a review of

recent studies (2001-2003) on human PCDD/PCDF exposure through the diet. Environmental

Research 2005; 97(1): 1-9.

Bocio A, Domingo JL, Falco G, Llobet JM. Concentrations of PCDD/PCDFs and PCBs in fish

and seafood from the Catalan (Spain) market: estimated human intake. Environment

International 2007; 33(2): 170-175.

Brouwer IA, Geelen A, Katan MB. n-3 Fatty acids, cardiac arrhythmia and fatal coronary

heart disease. Progress in Lipid Research 2006; 45(4): 357-367.

Brüders N, Knetsch G, Rappolder M. The German POP-DIOXIN DATABASE - A tool for

flexible evaluation available as webservice. Organohalogen Compounds 2005; 67: 2076-2078.

Brunsø K. Consumer research on fish in Europe. In: Luten JB, Oehlenschlager J, Olafsdottir G

(eds), Quality of fish from catch to consumer: labelling, monitoring and traceability. The

Netherlands: Wagheningen Academic Publishers, 2003: 335-344.

Brustad M, Sandanger T, Aksnes L, Lund E. Vitamin D status in a rural population of

northern Norway with high fish liver consumption. Public Health Nutrition 2004; 7(6): 783-

789.

Burdge GC, Wootton SA. Conversion of alpha-linolenic acid to eicosapentaenoic,

docosapentaenoic and docosahexaenoic acids in young women. British Journal of Nutrition

2002; 88(4): 411-420.

References

185

Burreau S, Zebuhr Y, Broman D, Ishaq R. Biomagnification of PBDEs and PCBs in food webs

from the Baltic Sea and the northern Atlantic Ocean. Science of the Total Environment 2006;

366(2-3): 659-672.

Cahu C, Salen P, de Lorgeril M. Farmed and wild fish in the prevention of cardiovascular

diseases: Assessing possible differences in lipid nutritional values. Nutrition Metabolism and

Cardiovascular Diseases 2004; 14(1): 34-41.

Calder PC. n-3 polyunsaturated fatty acids, inflammation, and inflammatory diseases.

American Journal of Clinical Nutrition 2006; 83(6): 1505S-1519S.

Cameron M, Van Staveren W. Manual on methodology for food consumption studies. New

York: Oxford University Press, 1988.

Candela M, Astiasaran I, Bello J. Effects of frying and warmholding on fatty acids and

cholesterol of sole (Solea solea), codfish (Gadus morhua) and hake (Merluccius merluccius). Food

Chemistry 1997; 58(3): 227-231.

Candela M, Astiasaran I, Bello J. Deep-fat frying modifies high-fat fish lipid fraction. Journal

of Agricultural and Food Chemistry 1998; 46(7): 2793-2796.

Capelli R, Drava G, Siccardi C, De Pellegrini R, Minganti V. Study of the distribution of trace

elements in six species of marine organisms of the Ligurian Sea (north-western

Mediterranean)--comparison with previous findings. Annali di Chimica 2004; 94(7-8): 533-546.

Caraher M, Coveney J. Public health nutrition and food policy. Public Health Nutrition 2004;

7(5): 591-598.

Carnovale E, Marletta L. Copmosizione degli alimenti - Aggiornamento 2000. Available at

http://www.inran.it. 2000.

Carrington CD, Bolger MP. An exposure assessment for methylmercury from seafood for

consumers in the United States. Risk Analysis 2002; 22(4): 689-699.

http://www.inran.it/

References

186

Carroll DN, Roth MT. Evidence for the cardioprotective effects of omega-3 fatty acids. The

Annals of pharmacotherapy 2002; 36(12): 1950-1956.

Caygill CPJ, Charlett A, Hill MJ. Fat, fish, fish oil and cancer. British Journal of Cancer 1996;

74(1): 159-164.

Chan HM, Egeland GM. Fish consumption, mercury exposure, and heart diseases. Nutrition

Reviews 2004; 62(2): 68-72.

Cibot C, Kolypczuk L. Seafood freshness and traceability for consumer safety. Sciences des

Aliments 2003; 23(4): 535-541.

Clarkson TW, Magos L. The toxicology of mercury and its chemical compounds. Critical

Reviews in Toxicology 2006; 36(8): 609-662.

Cohen JT, Bellinger DC, Connor WE, Kris-Etherton PM, Lawrence RS, Savitz DA, Shaywitz

BA, Teutsch SM, Gray GM. A quantitative risk-benefit analysis of changes in population fish

consumption. American Journal of Preventive Medicine 2005; 29(4): 325-334.

Commission of the European Communities. Nutrient and energy intakes for the European

Community. Report of the Scientific Committee for food. Luxembourg: Office for official

publications of the European Communities, 1993.

Copeman LA, Parrish CC. Lipids classes, fatty acids, and sterols in seafood from gilbert bay,

southern Labrador. Journal of Agricultural and Food Chemistry 2004; 52(15): 4872-4881.

Corsolini S, Ademollo N, Romeo T, Greco S, Focardi S. Persistent organic pollutants in edible

fish: a human and environmental health problem. Microchemical Journal 2005; 79(1-2): 115-

123.

Covaci A, Voorspoels S, Wyckmans J, Gelbin A, Neels H. Anthropogenic and natural

organohalogenated compounds in fish oil dietary supplements from various countries.

Organohalogen Compounds 2006; 68: 5-8.

References

187

Cunnane SC, Griffin BA. Nutrition and metabolism of lipids. In: Gibney MJ, Vorster HH,

Kok FJ (eds), Introduction to human nutrition. Oxford: Blackwell Science Ltd, 2002: 81-115.

Dahl L, Johansson L, Julshamn K, Meltzer HM. The iodine content of Norwegian foods and

diets. Public Health Nutrition 2004; 7(4): 569-576.

Danish Institute for Food and Veterinary Research. Danish Food Composition Databank

(revision 6.0). Available at http://www.foodcomp.dk. 2005.

Danish Veterinary and Food Administration. Combined report on dioxin in salmon and

herring form the Baltic Sea. Available at http://www.foedevarestyrelsen.dk. 2004.

Darnerud PO, Atuma S, Aune M, Bjerselius R, Glynn A, Grawe KP, Becker W. Dietary intake

estimations of organohalogen contaminants (dioxins, PCB, PBDE and chlorinated pesticides,

e.g. DDT) based on Swedish market basket data. Food and Chemical Toxicology 2006; 44(9):

1597-1606.

De Backer G. Nutrition and Health: an interuniversity study. Regional differences in dietary

habits, coronary risk factors and mortality rates in Belgium. I. Design and Methodology.

Acta Cardiologica 1984; 39: 285-292.

De Caterina R, Basta G. n-3 fatty acids and the inflammatory response - biological

background. European Heart Journal Supplements 2001; 3(D): D42-D49.

de Deckere EAM, Korver O, Verschuren PM, Katan MB. Health aspects of fish and n-3

polyunsaturated fatty acids from plant and marine origin. European Journal of Clinical

Nutrition 1998; 52(10): 749-753.

De Henauw S, Van Camp J, Sturtewagen G., Matthys C, Bilau M, Warnants N, Raes K, Van

Oeckel M, De Smet S. Simulated changes in fatty acid intake in humans through n-3 fatty

acid enrichment of foods from animal origin. Journal of the Science of Food and Agriculture

2007; 87: 200-211.

http://www.foodcomp.dk/

http://www.foedevarestyrelsen.dk/

References

188

De Vriese SR, De Henauw S, De Backer G, Dhont M, Christophe AB. Estimation of dietary fat

intake of Belgian pregnant women. Comparison of two methods. Annals of Nutrition and

Metabolism 2001; 45(6): 273-278.

De Vriese SR, Huybrechts I, Moreau M, Van Oyen H. De Belgische

Voedselconsumptiepeiling (The Belgian Food Consumption Survey). IPH/EPI REPORTS N°

2006 - 016 (http://www.iph.fgov.be/epidemio/epinl/index5.htm). Brussels: Scientific

Health Institute, Department of Epidemiology, 2006.

Diehl JM. Food preferences of 10- to 14-year-old boys and girls. Schweizerische Medinische

Wochenschrift 1999; 129: 151-161.

Din JN, Newby DE, Flapan AD. Science, medicine, and the future - Omega 3 fatty acids and

cardiovascular disease - fishing for a natural treatment. British Medical Journal 2004;

328(7430): 30-35.

Dodd KW, Guenther PM, Freedman LS, Subar AF, Kipnis V, Midthune D, Tooze JA, Krebs-

Smith SM. Statistical Methods for Estimating Usual Intake of Nutrients and Foods: A

Review of the Theory. Journal of the American Dietetic Association 2006; 106(10): 1640-1650.

Domergue F, Abbadi A, Heinz E. Relief for fish stocks: oceanic fatty acids in transgenic

oilseeds. Trends in Plant Science 2005; 10(3): 112-116.

Domingo JL, Bocio A, Falco G, Llobet JM. Benefits and risks of fish consumption: Part I. A

quantitative analysis of the intake of omega-3 fatty acids and chemical contaminants.

Toxicology 2007a; 230(2-3): 219-226.

Domingo JL, Bocio A, Marti-Cid R, Llobet JM. Benefits and risks of fish consumption: Part II.

RIBEPEIX, a computer program to optimize the balance between the intake of omega-3 fatty

acids and chemical contaminants. Toxicology 2007b; 230(2-3): 227-233.

Dyerberg J, Bang HO, Hjorne N. Fatty-Acid Composition of Plasma-Lipids in Greenland

Eskimos. American Journal of Clinical Nutrition 1975; 28(9): 958-966.

http://www.iph.fgov.be/epidemio/epinl/index5.htm)

References

189

Easton MDL, Luszniak D, der Geest E. Preliminary examination of contaminant loadings in

farmed salmon, wild salmon and commercial salmon feed. Chemosphere 2002; 46(7): 1053-

1074.

EFSA. Opinion of the scientific panel on contaminants in the food chain on a request from

the Commission related to mercury and methylmercury in food. The EFSA Journal 2004; 34:

1-14.

EFSA. Opinion of the scientific committee on a request from EFSA related to exposure

assessments. The EFSA Journal 2005a; 249: 1-26.

EFSA. Opinion of the scientific panel on contaminants in the food chain on a request from

the European Parliament related to the safety assessment of wild and farmed fish. The EFSA

Journal 2005b; 236: 1-118.

EFSA. Opinion of the scientific panel on contaminants in the food chain on a request of the

Commission related to the presence of non dioxin-like polychlorinated biphenyls (PCB) in

feed and food. The EFSA Journal 2005c; 284: 1-137.

EU Scientific Steering Committee. First report on the harmonisation of risk assessment

procedures in the European Union. 2000.

European Commission. COMMISSION REGULATION (EC) No 2065/2001 of 22 October 2001

laying down detailed rules for the application of Council Regulation (EC) No 104/2000 as regards

informing consumers about fishery and aquaculture products. Brussels: European Commission,

2001.

European Commission. REGULATION (EC) No 178/2002 of the European Parliament and of the

Council of 28 January 2002 laying down the general principles and requirements of food law,

establishing the European Food Safety Authority and laying down procedures in matters of food

safety. Brussels: European Commission, 2002.

European Commission. COMMISSION REGULATION (EC) No 1881/2006 of 19 December

2006 setting maximum levels for certain contaminants in foodstuffs. Brussels: European

Commission, 2006.

References

190

European Commission. Maritime Affairs. Available at http://ec.europa.eu/maritimeaffairs.

2007.

FAO. Traceability and labelling in fish trade. Report of the Ninth Session of the Committee

on Fisheries, Sub-Committee on Fish Trade, 10-14 February, Bremen, Germany. Rome:

FAO/COFI:FT, 2004.

FAO. The state of world fisheries and aquaculture 2006. Rome: FAO communication

division, 2007.

FAO/WHO. Report of the 27th Session Joint FAO/WHO Food Standards Programme, 28

June - 3 July, Geneva, Switzerland. Rome: Codex Alimentarius Commission, 2004.

Fattore E, Fanelli R, Turrini A, di Domenico A. Current dietary exposure to

polychlorodibenzo-p-dioxins, polychlorodibenzofurans, and dioxin-like

polychlorobiphenyls in Italy. Molecular Nutrition & Food Research 2006; 50(10): 915-921.

Favier J-C, Ireland-Ripert J, Toque C, Feinberg M. Répertoire général des aliments. Second

ed. Paris: INRA-AFSSA-CIQUAL-TEC & DOC, 1995.

Fernandes AR, Rose M, White S, Mortimer DN, Gem M. Dioxins and polychlorinated

biphenyls (PCBs) in fish oil dietary supplements: Occurrence and human exposure in the

UK. Food Additives and Contaminants 2006; 23(9): 939-947.

Food Standards Agency. McCance and Widdowson's. The composition of foods, Sixth

summary edition. Cambridge, UK: Royal Society of Chemistry, 2002.

Foran JA, Hites RA, Carpenter DO, Hamilton MC, Mathews-Amos A, Schwager SJ. A survey

of metals in tissues of farmed Atlantic and wild Pacific salmon. Environmental Toxicology and

Chemistry 2004; 23(9): 2108-2110.

Foran JA, Good DH, Carpenter DO, Hamilton MC, Knuth BA, Schwager SJ. Quantitative

Analysis of the Benefits and Risks of Consuming Farmed and Wild Salmon. Journal of

Nutrition 2005; 135(11): 2639-2643.

http://ec.europa.eu/maritimeaffairs

References

191

Forsyth DS, Casey V, Dabeka R, McKenzie AD. Methylmercury levels in predatory fish

species marketed in Canada. Food Additives and Contaminants 2004; 21(9): 849-856.

Frederiksen M, Gram L. Traceability. In: Huss HH, Ababouch L, Gram L (eds), Assessment

and management of seafood safety and quality. Rome: FAO, 2003.

Gall KL, Otwell WS, Koburger JA, Appledorf H. Effects of 4 Cooking Methods on the

Proximate, Mineral and Fatty-Acid Composition of Fish Fillets. Journal of Food Science 1983;

48(4): 1068-1074.

Gebauer SK, Psota TL, Harris WS, Kris-Etherton PM. n-3 Fatty acid dietary

recommendations and food sources to achieve essentiality and cardiovascular benefits.

American Journal of Clinical Nutrition 2006; 83(6): S1526-S1535.

Gibney MJ, van der Voet H. Introduction to the Monte Carlo project and the approach to the

validation of probabilistic models of dietary exposure to selected food chemicals. Food

Additives and Contaminants 2003; 20: S1-S7.

Gilsenan MB, Lambe J, Gibney MJ. Assessment of food intake input distributions for use in

probabilistic exposure assessments of food additives. Food Additives and Contaminants 2003;

20(11): 1023-1033.

Gokoglu N, Yerlikaya P, Cengiz E. Effects of cooking methods on the proximate composition

and mineral contents of rainbow trout (Oncorhynchus mykiss). Food Chemistry 2004; 84(1): 19-

22.

Greenfield H, Southgate DAT. Food composition data: Production, management and use.

Second ed. Rome: FAO, 2003.

Gross T. Consumer attitudes towards health and food safety. In: Luten JB, Oehlenschlager J,

Olafsdottir G (eds), Quality of fish from catch to consumer: labelling, monitoring and traceability.

The Netherlands: Wageningen Academic Publishers, 2003: 401-411.

Haard NF. Control of Chemical-Composition and Food Quality Attributes of Cultured Fish.

Food Research International 1992; 25(4): 289-307.

References

192

Hamre K, Lie O, Sandnes K. Seasonal development of nutrient composition, lipid oxidation

and colour of fillets from Norwegian spring-spawning herring (Clupea harengus L.). Food

Chemistry 2003; 82(3): 441-446.

Harel Z, Riggs S, Vaz R, White L, Menzies G. Omega-3 polyunsaturated fatty acids in

adolescents: Knowledge and consumption. Journal of Adolescent Health 2001; 28(1): 10-15.

Harris WS. n-3 Fatty acid fortification: opportunities and obstacles. British Journal of Nutrition

2007; 97: 593-595.

Harrison RA, Sagara M, Rajpura A, Armitage L, Birt N, Birt CA, Yamori Y. Can foods with

added soya-protein or fish-oil reduce risk factors for coronary disease? A factorial

randomised controlled trial. Nutrition, Metabolism and Cardiovascular Diseases 2004; 14(6): 344-

350.

Hathcock JN, Shao A, Vieth R, Heaney R. Risk assessment for vitamin D. American Journal of

Clinical Nutrition 2007; 85(1): 6-18.

Health Canada. Canadian Nutrient File, version 2005. Available at http://www.hc-

sc.gc.ca/food-aliment/ns-sc/nr-rn/surveillance/cnf-fcen/e_index.html. 2005.

Hepburn FN, Exler J, Weihrauch JL. Provisional Tables on the Content of Omega-3-Fatty-

Acids and Other Fat Components of Selected Foods. Journal of the American Dietetic

Association 1986; 86(6): 788-793.

Hibbeln JR. Fish consumption and major depression. Lancet 1998; 351(9110): 1213.

Hites RA, Foran JA, Carpenter DO, Hamilton MC, Knuth BA, Schwager SJ. Global

assessment of organic contaminants in farmed salmon. Science 2004a; 303(5655): 226-229.

Hites RA, Foran JA, Schwager SJ, Knuth BA, Hamilton MC, Carpenter DO. Global

assessment of polybrominated diphenyl ethers in farmed and wild salmon. Environmental

Science & Technology 2004b; 38(19): 4945-4949.

http://www.hc-sc.gc.ca/food-aliment/ns-sc/nr-rn/surveillance/cnf-fcen/e_index.html

http://www.hc-sc.gc.ca/food-aliment/ns-sc/nr-rn/surveillance/cnf-fcen/e_index.html

References

193

Holden JM, Bhagwat SA, Patterson KY. Development of a multi-nutrient data quality

evaluation system. Journal of Food Composition and Analysis 2002; 15(4): 339-348.

Holland B, Brown J, Buss DH. Fish and fish products. The third supplement to McCance &

Widdowson's. The composition of food. Fifth ed. Cambridge: The Royal Society of

Chemistry, 1993.

Hollis BW, Wagner CL. Assessment of dietary vitamin D requirements during pregnancy

and lactation. American Journal of Clinical Nutrition 2004; 79(5): 717-726.

Honkanen P, Brunsø K. On the average European fish consumption is below recommended

levels. Deliverable 4, Project 2.1, SEAFOODplus. Available at

http://www.seafoodplus.org/Europen_fish_consumption.411.0.html. 2007.

Hooper L, Thompson RL, Harrison RA, Summerbell CD, Ness AR, Moore HJ, Worthington

HV, Durrington PN, Higgins JPT, Capps NE, Riemersma RA, Ebrahim SBJ, Smith GD. Risks

and benefits of omega 3 fats for mortality, cardiovascular disease, and cancer: systematic

review. British Medical Journal 2006; 332(7544): 752-760.

Howe P, Meyer B, Record S, Baghurst K. Dietary intake of long-chain [omega]-3

polyunsaturated fatty acids: contribution of meat sources. Nutrition 2006; 22(1): 47-53.

Hursting SD, Thornquist M, Henderson MM. Types of Dietary-Fat and the Incidence of

Cancer at 5 Sites. Preventive Medicine 1990; 19(3): 242-253.

Huybrechts I, De Bacquer D, Matthys C, De Backer G, De Henauw S. Validity and

reproducibility of a semi-quantitative food-frequency questionnaire for estimating calcium

intake in Belgian preschool children. British Journal of Nutrition 2006; 95(4): 802-816.

Huybrechts I, De Henauw S. Energy and nutrient intakes by pre-school children in Flanders-

Belgium. British Journal of Nutrition 2007; 20: 1-11.

Innis SM, Elias SL. Intakes of essential n-6 and n-3 polyunsaturated fatty acids among

pregnant Canadian women. American Journal of Clinical Nutrition 2003; 77(2): 473-478.

http://www.seafoodplus.org/Europen_fish_consumption.411.0.html

References

194

Innis SM, Vaghri Z, King DJ. N-6 Docosapentaenoic acid is not a predictor of low

docosahexaenoic acid status in Canadian preschool children. American Journal of Clinical

Nutrition 2004; 80(3): 768-773.

Innis SM. Dietary (n-3) Fatty Acids and Brain Development. Journal of Nutrition 2007; 137(4):

855-859.

Institut Paul Lambin. Table de composition des aliments. Buxelles: Institut Paul Lambin,

2004.

Institute of Medicine. Dietary reference intakes for calcium, phosphorus, magnesium,

vitamin D, and fluoride. Washington: The National Academies Press, 1997.

Institute of Medicine. Dietary Reference Intakes for Vitamin A, Vitamin K, Arsenic, Boron,

Chromium, Copper, Iodine, Iron, Manganese, Molybdenum, Nickel, Silicon, Vanadium, and

Zinc. Washington: The National Academies Press, 2001.

Institute of Medicine. Dietary reference intakes for energy, carbohydrate, fiber, fat, fatty

acids, cholesterol, protein, and amino acids. Washington: The National Academies Press,

2005.

Iowa State University. C-side. 2006.

Isosaari P, Kiviranta H, Lie O, Lundebye AK, Ritchie G, Vartiainen T. Accumulation and

distribution of polychlorinated dibenzo-p-dioxin, dibenzofuran, and polychlorinated

biphenyl congeners in Atlantic salmon (Salmo salar). Environmental Toxicology and Chemistry

2004; 23(7): 1672-1679.

ISSFAL. Report of the sub-committee on the recommendations of polyunsaturated fatty acids

in healthy adults. Available at http://www.issfal.org. 2007.

Jacobs MN, Covaci A, Gheorghe A, Schepens P. Time trend investigation of PCBs, PBDEs,

and organochlorine pesticides in selected n-3 polyunsaturated fatty acid rich dietary fish oil

and vegetable oil supplements; Nutritional relevance for human essential n-3 fatty acid

requirements. Journal of Agricultural and Food Chemistry 2004; 52(6): 1780-1788.

http://www.issfal.org/

References

195

Johansson LRK, Solvoll K, Bjorneboe GEA, Drevon CA. Intake of very-long-chain n-3 fatty

acids related to social status and lifestyle. European Journal of Clinical Nutrition 1998; 52(10):

716-721.

Jooken K, Lauryssen S. Aquacultuur en etikettering van vis (Aquaculture and labelling of

fish). Test-Aankoop 2006; 496: 28-32.

Judd N, O'Neill SM, Kalman DA. Are seafood PCB data sufficient to assess health risk for

high seafood consumption groups? Human and Ecological Risk Assessment 2003a; 9(3): 691-

707.

Judd NL, Karr JR, Griffith WC, Faustman EM. Challenges in defining background levels for

human and ecological risk assessments. Human and Ecological Risk Assessment 2003b; 9: 1623-

1632.

Juresa D, Blanusa M. Mercury, arsenic, lead and cadmium in fish and shellfish from the

Adriatic Sea. Food Additives and Contaminants 2003; 20(3): 241-246.

Kaizer L, Boyd NF, Kriukov V, Tritchler D. Fish Consumption and Breast-Cancer Risk - An

Ecological Study. Nutrition and Cancer-An International Journal 1989; 12(1): 61-68.

Karl H, Munkner W, Krause S, Bagge I. Determination, spatial variation and distribution of

iodine in fish. Deutsche Lebensmittel-Rundschau 2001; 97(3): 89-96.

Karl H, Ruoff U, Bluthgen A. Levels of dioxins in fish and fishery products on the German

market. Chemosphere 2002; 49(7): 765-773.

Karl H, Kuhlmann H, Ruoff U. Transfer of PCDDs and PCDFs into the edible parts of farmed

rainbow trout, Oncorhynchus mykiss (Walbaum), via feed. Aquaculture Research 2003; 34(12):

1009-1014.

Karlsen KM, Senneset G. Traceability: Simulated recall of fish products. In: Luten JB,

Jacobsen C, Bekaert K, Saebo A, Oehlenschlager J (eds), Seafood research from fish to dish.

Quality, safety and processing of wild and farmed fish. Wageningen: Wageningen Academic

Publisher, 2006: 251-261.

References

196

Kelly J, Turner JJ, McKenna K. What parents think: children and healthy eating. British Food

Journal 2006; 108(5): 413-423.

Kinney AJ. Metabolic engineering in plants for human health and nutrition. Current Opinion

in Biotechnology 2006; 17(2): 130-138.

Kinsella JE, Lokesh B, Stone RA. Dietary N-3 Polyunsaturated Fatty-Acids and Amelioration

of Cardiovascular-Disease - Possible Mechanisms. American Journal of Clinical Nutrition 1990;

52(1): 1-28.

Kiviranta H, Ovaskainen ML, Vartiainen T. Market basket study on dietary intake of

PCDD/Fs, PCBs, and PBDEs in Finland. Environment International 2004; 30(7): 923-932.

Knowles TG, Farrington D, Kestin SC. Mercury in UK imported fish and shellfish and UK-

farmed fish and their products. Food Additives and Contaminants 2003; 20(9): 813-818.

Kodavanti PR, Kannan N, Yamashita N, Derr-Yellin EC, Ward TR, Burgin DE, Tilson HA,

Birnbaum LS. Differential effects of two lots of aroclor 1254: congener-specific analysis and

neurochemical end points. Environmental Health Perspectives 2001; 109(11): 1153-1161.

Kornitzer M, Dramaix M. The Belgian Interuniversity Research on Nutrition and Health

(B.I.R.N.H.): general introduction. For the B.I.R.N.H. Study Group. Acta Cardiologica 1989;

44(2): 89-99.

Korver O, Katan MB. The elimination of trans fats from spreads: How science helped to turn

an industry around. Nutrition Reviews 2006; 64(6): 275-279.

Krebs-Smith SM, Kott PS, Guenther PM. Mean proportion and population proportion: two

answers to the same question? Journal of the American Dietetic Association 1989; 89(5): 671-676.

Kris-Etherton PM, Harris WS, Appel LJ. Fish consumption, fish oil, omega-3 fatty acids, and

cardiovascular disease. Circulation 2002; 106(21): 2747-2757.

References

197

Kris-Etherton PM, Harris WS, Appel LJ. Omega-3 fatty acids and cardiovascular disease -

New recommendations from the American Heart Association. Arteriosclerosis Thrombosis and

Vascular Biology 2003; 23(2): 151-152.

Kroes R, Muller D, Lambe J, Lowik MRH, van Klaveren J, Kleiner J, Massey R, Mayer S,

Urieta I, Verger P, Visconti A. Assessment of intake from the diet. Food and Chemical

Toxicology 2002; 40(2-3): 327-385.

Kromhout D. 'Protective nutrients' and up-to-date dietary recommendations. European Heart

Journal Supplements 2001; 3(D): D33-D36.

Lambe J. The use of food consumption data in assessments of exposure to food chemicals

including the application of probabilistic modelling. Proceedings of the Nutrition Society 2002;

61(1): 11-18.

Lamberg-Allardt C. Vitamin D in foods and as supplements. Progress in Biophysics and

Molecular Biology 2006; 92(1): 33-38.

Lands WEM. Fish, omega-3 and human health. Second ed. Illinois: AOCS Press, 2005.

Larsson SC, Kumlin M, Ingelman-Sundberg M, Wolk A. Dietary long-chain n-3 fatty acids

for the prevention of cancer: a review of potential mechanisms. American Journal of Clinical

Nutrition 2004; 79: 935-945.

Lawrance P. Speciation, bioavailability, biopotency and functional markers for D vitamers.

FA50/2002/053/W063-000. 2002.

Lawrence M, Rayner M. Functional foods and health claims: a public health policy

perspective. Public Health Nutrition 1998; 1(2): 75-82.

Leek S, Maddock S, Foxall G. Situational determinants of fish consumption. British Food

Journal 2000; 102(1): 18-39.

Legrand P, Bourre JM, Descomps B, Durand G, Renaud S. Lipides. In: Martin A (ed), Apports

nutritionnels conseillés pour la population française. Paris: Tec & Doc, 2001: 63-82.

References

198

Linseisen J, Schulze MB, Saadatian-Elahi M, Kroke A, Miller AB, Boeing H. Quantity and

quality of dietary fat, carbohydrate, and fiber intake in the German EPIC cohorts. Annals of

Nutrition and Metabolism 2003; 47(1): 37-46.

Liou YA, King DJ, Zibrik D, Innis SM. Decreasing Linoleic Acid with Constant {alpha}-

Linolenic Acid in Dietary Fats Increases (n-3) Eicosapentaenoic Acid in Plasma

Phospholipids in Healthy Men. Journal of Nutrition 2007; 137(4): 945-952.

Liu M, Wallin R, Saldeen T. Effect of bread containing stable fish oil on plasma phospholipid

fatty acids, triglycerides, HDL-cholesterol, and malondialdehyde in subjects with

hyperlipidemia. Nutrition Research 2001; 21(11): 1403-1410.

Loosemore ED, Judge MP, Lammi-Keefe CJ. Dietary intake of essential and long-chain

polyunsaturated fatty acids in pregnancy. Lipids 2004; 39(5): 421-424.

Lopez A, Rueda C, Armentia A, Rodriguez M, Cuervo L, Ocio JA. Validation and sensitivity

analysis of a probabilistic model for dietary exposure assessment to pesticide residues with

a Basque Country duplicate diet study. Food Additives and Contaminants 2003; 20: S87-S101.

Lupin H. Traceability. Document presented at the Tenth Session of Committees on Fisheries,

Sub-Committee on Fish Trade, 30 May-2 June, Santiago de Compostella, Spain. 2006.

Mai J, Shimp J, Weihrauch J, Kinsella JE. Lipids of Fish Fillets - Changes Following Cooking

by Different Methods. Journal of Food Science 1978; 43(6): 1669-1674.

Mann NJ, Johnson LG, Warrick GE, Sinclair AJ. The Arachidonic-Acid Content of the

Australian Diet Is Lower Than Previously Estimated. Journal of Nutrition 1995; 125(10): 2528-

2535.

Mann NJ, Sinclair AJ, Percival P, Lewis JL, Meyer BJ, Howe PR. Development of a database

of fatty acids in Australian foods. Nutrition & Dietetics 2003; 60: 42-45.

Margetts B. An overview of public health nutrition. In: Gibney MJ, Margetts B, Kearney J,

Arab L (eds), Public Health Nutrition. Oxford, UK: Blackwell Publishing, 2004: 1-25.

References

199

Marien K. The importance of weight-normalized exposure data when issuing fish advisories

for protection of public health. Environmental Health Perspectives 2002; 110(7): 671-677.

Matthys C, Bilau M, Govaert Y, Moons E, De Henauw S, Willems JL. Risk assessment of

dietary acrylamide intake in Flemish adolescents. Food and Chemical Toxicology 2005; 43(2):

271-278.

Matthys C, De Henauw S, Bellemans M, De Maeyer M, De Backer G. Sources of saturated

fatty acids in Belgian adolescents' diet: implications for the development of food-based

dietary guidelines. British Journal of Nutrition 2006; 95(3): 546-554.

Matthys C, De Henauw S, Devos C, De Backer G. Estimated energy intake, macronutrient

intake and meal pattern of Flemish adolescents. European Journal of Clinical Nutrition 2003;

57(2): 366-375.

Mattila P. Analysis of cholecalciferol, ergocalciferol and their 25-hydroxylated metabolites in

food by HPLC. University of Helsinki, Department of Applied Chemistry and Microbiology,

1995.

Mattila P, Piironen V, Haapala R, Hirvi T, UusiRauva E. Possible factors responsible for the

high variation in the cholecalciferol contents of fish. Journal of Agricultural and Food Chemistry

1997; 45(10): 3891-3896.

Mattila P, Piironen V, Uusi-Rauva E, Koivistoinen P. Cholecalciferol and 25-

hydroxycholecalciferol contents in fish and fish products. Journal of Food Composition and

Analysis 1995; 8: 232-243.

Mattila P, Ronkainen R, Lehikoinen K, Piironen V. Effect of household cooking on the

vitamin D content in fish, eggs, and wild mushrooms. Journal of Food Composition and

Analysis 1999; 12: 153-160.

Mattila PH, Piironen VI, UusiRauva EJ, Koivistoinen PE. New analytical aspects of vitamin D

in foods. Food Chemistry 1996; 57(1): 95-99.

References

200

McMichael AJ, Butler CD. Fish, Health, and Sustainability. American Journal of Preventive

Medicine 2005; 29(4): 322-323.

Metcalf RG, James MJ, Mantzioris E, Cleland LG. A practical approach to increasing intakes

of n-3 polyunsaturated fatty acids: use of novel foods enriched with n-3 fats. European

Journal of Clinical Nutrition 2003; 57(12): 1605-1612.

Meyer BJ, Mann NJ, Lewis JL, Milligan GC, Sinclair AJ, Howe PRC. Dietary intakes and food

sources of omega-6 and omega-3 polyunsaturated fatty acids. Lipids 2003; 38(4): 391-398.

Mozaffarian D, Rimm EB. Fish Intake, Contaminants, and Human Health: Evaluating the

Risks and the Benefits. Journal of the American Medical Association 2006; 296(15): 1885-1899.

Mullin GE, Dobs A. Vitamin D and Its Role in Cancer and Immunity: A Prescription for

Sunlight. NCP - Nutrition in Clinical Practice 2007; 22(3): 305-322.

Murphy KJ, Meyer BJ, Mori TA, Burke V, Mansour J, Patch CS, Tapsell LC, Noakes M,

Clifton PM, Barden A, Pudey IB, Beilin LJ, Howe PR. Impact of foods enriched with n-3

long-chain polyunsaturated fatty acids on erythrocyte n-3 levels and cardiovascular risk

factors. British Journal of Nutrition 2007; 97: 749-757.

Muskiet FAJ, van Goor SA, Kuipers RS, Velzing-Aarts FV, Smit EN, Bouwstra H, Janneke

Dijck-Brouwer DA, Rudy Boersma E, Hadders-Algra M. Long-chain polyunsaturated fatty

acids in maternal and infant nutrition. Prostaglandins Leukotrienes and Essential Fatty Acids

2006; 75(3): 135-144.

Napier JA, Sayanova O. The production of very-long-chain PUFA biosynthesis in transgenic

plants: towards a sustainable source of fish oils. Proceedings of the Nutrition Society 2005;

64(3): 387-393.

National Public Health Institute of Finland. Finnish Food Composition Database. Available

at http://www.fineli.fi. 2004.

http://www.fineli.fi/

References

201

Naylor RL, Goldburg RJ, Primavera JH, Kautsky N, Beveridge MCM, Clay J, Folke C,

Lubchenco J, Mooney H, Troell M. Effect of aquaculture on world fish supplies. Nature 2000;

405(6790): 1017-1024.

Nelson M, Black AE, Morris JA, Cole TJ. Between- and within-subject variation in nutrient

intake from infancy to old age: estimating the number of days required to rank dietary

intakes with desired precision. American Journal of Clinical Nutrition 1989; 50(1): 155-167.

Nettleton JA, Exler J. Nutrients in Wild and Farmed Fish and Shellfish. Journal of Food Science

1992; 57(2): 257-260.

NEVO Foundation. Nederlands Voedingsstoffenbestand 2001 (Dutch Food Composition

Table 2001). Den Haag: Nederlands Voedingscentrum, 2001.

NEVO Foundation. Nederlands Voedingsstoffenbestand 2006 (Dutch Food Composition

Table 2006). Den Haag: Nederlands Voedingscentrum, 2006.

Nierentz J. The growth of aquaculture: A global perspective. Paper presented at: Future

Aquaculture conference, 14-15 September, Buino, Italy. 2006.

NIFES. Nifes Seafood-data - Levels of undesirable substances. Available at

http://www.nifes.no/seafood-data/indexe.html. 2004.

NUBEL. Belgische voedingsmiddelentabel. Derde editie. (Belgian food composition table.

Third edition). Third ed. Brussels: Ministerie van Volksgezondheid, 1999.

Nusser SM, Carriquiry AL, Dodd KW, Fuller WA. A semiparametric transformation

approach to estimating usual daily intake distributions. Journal of the American Statistical

Association 1996; 61: 1440-1449.

Olsen SO, Scholderer J, Brunsø K, Verbeke W. Exploring the relationship between

convenience and fish consumption: A cross-cultural study. Appetite 2007; 49: 84-91.

http://www.nifes.no/seafood-data/indexe.html

References

202

Olsson GB, Olsen RL, Carlehog M, Ofstad R. Seasonal variations in chemical and sensory

characteristics of farmed and wild Atlantic halibut (Hippoglossus hippoglossus). Aquaculture

2003; 217(1-4): 191-205.

Otto SJ, van Houwelingen AC, Badart-Smook A, Hornstra G. Changes in the maternal

essential fatty acid profile during early pregnancy and the relation of the profile to diet.

American Journal of Clinical Nutrition 2001; 73(2): 302-307.

Ovesen L, Andersen R, Jakobsen J. Geographical differences in vitamin D status, with

particular reference to European countries. Proceedings of the Nutrition Society 2003; 62: 813-

821.

Palisade Corporation. Bestfit Distribution Fitting for Windows, Version 4.5. Newfield, NY

USA:2002.

Pàlsson PG, Storoy J, Frederiksen M, Olsen P. Traceability and electronic transmission of

qualitative data for fish products. Status report nr.3. Tromso, Norway: The Nordic Minister

Council, 2000.

Patch CS, Tapsell LC, Mori TA, Meyer BJ, Murphy KJ, Mansour J, Noakes M, Clifton PM,

Pudey IB, Beilin LJ, Annison G, Howe PR. The use of novel foods enriched with long-chain

n-3 fatty acids to increase dietary intake: a comparison of methodologies assessing intake.

Journal of the American Dietetic Association 2005a; 105: 1918-1926.

Patch CS, Tapsell LC, Williams PG. Attitudes and intentions toward purchasing novel food

enriched with omega-3 fatty acids. Journal of Nutrition Education and Behavior 2005b; 37: 235-

241.

Paulo MJ, van der Voet H, Wood JC, Marion GR, van Klaveren JD. Analysis of multivariate

extreme intakes of food chemicals. Food and Chemical Toxicology 2006; 44(7): 994-1005.

Petersen BJ. Probabilistic modelling: theory and practice. Food Additives and Contaminants

2000; 17(7): 591-599.

References

203

Pieniak Z, Verbeke W, Scholderer J, Brunsø K, Olsen SO. European consumers' use of and

trust in information sources about fish. Food Quality and Preference 2007; 18(8): 1050-1063.

Piironen V, Toivo J, Lampi AM. New data for cholesterol contents in meat, fish, milk, eggs

and their products consumed in Finland. Journal of Food Composition and Analysis 2002; 15(6):

705-713.

Plessi M, Bertelli D, Monzani J. Mercury and selenium content in selected seafood. Journal of

Food Composition and Analysis 2001; 14: 461-467.

Ponce RA, Bartell SM, Wong EY, LaFlamme D, Carrington C, Lee RC, Patrick DL, Faustman

EM, Bolger M. Use of quality-adjusted life year weights with dose-response models for

public health decisions: a case study of the risks and benefits of fish consumption. Risk

Analysis 2000; 20(4): 529-542.

Prather RS. Cloned transgenic heart-healthy pork? Transgenic Research 2006; 15(4): 405-407.

Pynaert I, Matthys C, Bellemans M, De Maeyer M, De Henauw S, De Backer G. Iron intake

and dietary sources of iron in Flemish adolescents. European Journal of Clinical Nutrition 2005;

59(7): 826-834.

R Development Core Team. R: A language and environment for statistical computing. R

Foundation for Statistical Computing, Vienna, Austria. ISBN 3-900051-07-0, URL

http://www.R-project.org. 2006.

Raemaeckers M. Contamination data ICES [Personal communication]. 2005.

Raes K, De Smet S, Demeyer D. Effect of dietary fatty acids on incorporation of long chain

polyunsaturated fatty acids and conjugated linoleic acid in lamb, beef and pork meat: a

review. Animal Feed Science and Technology 2004; 113(1-4): 199-221.

Ratledge C. Fatty acid biosynthesis in microorganisms being used for Single Cell Oil

production. Biochimie 2004; 86(11): 807-815.

http://www.r-project.org/

References

204

Reinert A, Rohrmann S, Becker N, Linseisen J. Lifestyle and diet in people using dietary

supplements : A German cohort study. European Journal of Nutrition 2007; 46(3): 165-173.

Richardson AJ. Clinical trials of fatty acid treatment in ADHD, dyslexia, dyspraxia and the

autistic spectrum. Prostaglandins Leukotrienes and Essential Fatty Acids 2004a; 70(4): 383-390.

Richardson AJ. Long-chain polyunsaturated fatty acids in childhood developmental and

psychiatric disorders. Lipids 2004b; 39(12): 1215-1222.

Richardson AJ, Montgomery P. The Oxford-Durham Study: A Randomized, Controlled Trial

of Dietary Supplementation With Fatty Acids in Children With Developmental

Coordination Disorder. Pediatrics 2005; 115(5): 1360-1366.

Rubingh CM, Kruizinga AG, Hulshof KFAM, Brussaard JH. Validation and sensitivity

analysis of probabilistic models of dietary exposure to micronutrients: an example based on

vitamin B-6. Food Additives and Contaminants 2003; 20: S50-S60.

Ruxton C. Health benefits of omega-3 fatty acids. Nursing Standard 2004; 18(48): 38-42.

Ruxton CH, Reed SC, Simpson MJ, Millington KJ. The health benefits of omega-3

polyunsaturated fatty acids: a review of the evidence. Journal of Human Nutrition and

Dietetics 2004; 17(5): 449-459.

Ruxton CHS, Calder PC, Reed SC, Simpson MJA. The impact of long-chain n-3

polyunsaturated fatty acids on human health. Nutrition Research Reviews 2005; 18(1): 113-129.

SACN/COT. Scientific Advisory Committee on Nutrition (SACN) and Committee on

Toxicity (COT), Advice on fish consumption: benefits and risks. Available at

http://www.sacn.gov.uk. Norwich: TSO, 2004.

Salama AA, Mohamed MAM, Duval B, Potter TL, Levin RE. Polychlorinated biphenyl

concentration in raw and cooked North Atlantic bluefish (Pomatomus saltatrix) fillets. Journal

of Agricultural and Food Chemistry 1998; 46(4): 1359-1362.

http://www.sacn.gov.uk/

References

205

Salem NJ, Pawlosky RJ. Docosahexaenoic acid is an essential nutrient in the nervous system.

Journal of Nutritional Science and Vitaminology 1992; Spec No: 153-156.

Salo-Vaananen P, Ollilainen V, Mattila P, Lehikoinen K, Salmela-Molsa E, Piironen V.

Simultaneous HPLC analysis of fat-soluble vitamins in selected animal products after small-

scale extraction. Food Chemistry 2000; 71(4): 535-543.

Salonen JT, Seppanen K, Lakka TA, Salonen R, Kaplan GA. Mercury accumulation and

accelerated progression of carotid atherosclerosis: a population-based prospective 4-year

follow-up study in men in eastern Finland. Atherosclerosis 2000; 148(2): 265-273.

Salvini S, Parpinel M, Gnagnarella P, Maisonneuve P, Turrini A. Food Composition Database

for Epidemiological Studies in Italy. 1998.

Sanchez-Muniz FJ, Viejo JM, Medina R. Deep-frying of sardines in different culinary fats.

Changes in the fatty acid composition of sardines and frying fats. Journal of Agricultural and

Food Chemistry 1992; 40: 2252-2256.

Sanzo JM, Dorronsoro M, Amiano P, Amurrio A, Aguinagalde FX, Azpiri MA. Estimation

and validation of mercury intake associated with fish consumption in an EPIC cohort of

Spain. Public Health Nutrition 2001; 4(5): 981-988.

Sasaki S, Horacsek M, Kesteloot H. An Ecological Study of the Relationship Between Dietary-

Fat Intake and Breast-Cancer Mortality. Preventive Medicine 1993; 22(2): 187-202.

Scientific Committee on Food. Opinion of the SCF on the risk assessment of dioxins and

dioxin-like PCBs in food. Update based on new scientific information available since

adoption of the SCF opinion of 22nd November 2000 . CS/CNTM/DIOXIN/20 final.

Available at http://ec.europa.eu/food/fs/sc/scf/out90_en.pdf. 2001.

Scientific Committee on Food. Opinion of the Scientific Committee on Food on the Tolerable

Upper Intake Level of vitamin D. Brussel, Belgium: European Commission, 2002.

Scientific Institute of Public Health. Belgian Health Interview Survey, 2004 - Interactive

Analyses. Available at http://www.iph.fgov.be. 2006.

http://ec.europa.eu/food/fs/sc/scf/out90_en.pdf

http://www.iph.fgov.be/

References

206

Sérot T, Gandemer G, Demaimay M. Lipid and fatty acid composition of muscle from farmed

and wild adult turbot. Aquaculture International 1998; 6: 331-343.

Sichert-Hellert W, Wenz G, Kersting M. Vitamin Intakes from Supplements and Fortified

Food in German Children and Adolescents: Results from the DONALD Study. Journal of

Nutrition 2006; 136(5): 1329-1333.

Sidhu KS. Health benefits and potential risks related to consumption of fish or fish oil.

Regulatory Toxicology and Pharmacology 2003; 38(3): 336-344.

Simm M, Roots O, Kotta J, Lankov A, Henkelmann B, Shen H, Schramm KW. PCDD/Fs in

sprat (Sprattus sprattus balticus) from the Gulf of Finland, the Baltic Sea. Chemosphere 2006;

65(9): 1570-1575.

Simopoulos AP. Summary of the Nato Advanced Research Workshop on Dietary Omega-3

and Omega-6 Fatty-Acids - Biological Effects and Nutritional Essentiality. Journal of

Nutrition 1989; 119(4): 521-528.

Simopoulos AP. Evolutionary aspects of omega-3 fatty acids in the food supply.

Prostaglandins Leukotrienes and Essential Fatty Acids 1999; 60(5-6): 421-429.

Simopoulos AP. The importance of the ratio of omega-6/omega-3 essential fatty acids.

Biomedicine & Pharmacotherapy 2002; 56(8): 365-379.

Simopoulos AP, Leaf A, Salem N. Workshop Statement on the Essentiality of and

Recommended Dietary Intakes for Omega-6 and Omega-3 Fatty Acids. Prostaglandins

Leukotrienes and Essential Fatty Acids 2000; 63(3): 119-121.

Sioen IA, Pynaert I, Matthys C, De Backer G, Van Camp J, De Henauw S. Dietary intakes and

food sources of fatty acids for Belgian women, focused on n-6 and n-3 polyunsaturated fatty

acids. Lipids 2006a; 41(5): 415-422.

Sioen I, Haak L, Raes K, Hermans C, De Henauw S, De Smet S, Van Camp J. Effects of pan-

frying in margarine and olive oil on the fatty acid composition of cod and salmon. Food

Chemistry 2006b; 98(4): 609-617.

References

207

Sirot V, Tard A, Marchand P, Le Bizec B, Venisseau A, Brosseau A., Volatier JL, Leblanc JC.

Food exposure to persistent organic pollutants among French high seafood consumers

(Calipso study). Organohalogen Compounds 2006; 68: 383-386.

Slob W. Modeling long-term exposure of the whole population to chemicals in food. Risk

Analysis 1993; 13(5): 525-530.

Solomon K, Giesy J, Jones P. Probabilistic risk assessment of agrochemicals in the

environment. Crop Protection 2000; 19(8-10): 649-655.

Sondergaard H, Leerbeck E. Vitamin D content of Danish foods. Supplement concerning fish.

Soborg, Denmark: 1984.

Sontrop J, Campbell MK. [omega]-3 polyunsaturated fatty acids and depression: A review of

the evidence and a methodological critique. Preventive Medicine 2006; 42(1): 4-13.

Souci SW, Fachmann W, Kraut H. Food composition and nutrition tables. Sixth ed. München:

Scientific Publishers Stuttgart, 2000.

Stolyhwo A, Kolodziejska I, Sikorski ZE. Long chain polyunsaturated fatty acids in smoked

Atlantic mackerel and Baltic sprats. Food Chemistry 2006; 94(4): 589-595.

Storelli MM, Giacominelli-Stuffler R, Storelli A, D'Addabbo R, Palermo C, Marcotrigiano GO.

Survey of total mercury and methylmercury levels in edible fish from the Adriatic Sea. Food

Additives and Contaminants 2003a; 20(12): 1114-1119.

Storelli MM, Giacominelli-Stuffler R, Storelli A, Marcotrigiano GO. Polychlorinated

biphenyls in seafood: contamination levels and human dietary exposure. Food Chemistry

2003b; 82(3): 491-496.

Storelli MM, Storelli A, Giacominelli-Stuffler R, Marcotrigiano GO. Mercury speciation in the

muscle of two commercially important fish, hake (Merluccius merluccius) and striped mullet

(Mullus barbatus) from the Mediterranean sea: estimated weekly intake. Food Chemistry 2005;

89: 295-300.

References

208

Storelli MM, Storelli A, Marcotrigiano GO. Polychlorinated biphenyls, hexachlorobenzene,

hexachlorocyclohexane isomers, and pesticide organochlorine residues in cod-liver oil

dietary supplements. Journal of Food Protection 2004; 67(8): 1787-1791.

Storelli MM, Stuffler RG, Marcotrigiano GO. Total and methylmercury residues in tuna-fish

from the Mediterranean sea. Food Additives and Contaminants 2002; 19(8): 715-720.

Storelli MM, Stuffler RG, Storelli A, Marcotrigiano GO. Total mercury and methylmercury

content in edible fish from the Mediterranean Sea. Journal of Food Protection 2003c; 66(2): 300-

303.

Strain J, Cashman K. Minerals and Trace Elements. In: Gibney MJ, Vorster HH, Kok FJ (eds),

Introduction to human nutrition. Oxford: Blackwell Science Ltd, 2002: 177-224.

Sugiyama Jogakuen University. Food Composition Database in Sugiyama Univ. Standard

Tables of Food Composition in Japan. Available at http://database.food.sugiyama-

u.ac.jp/index_asia.php. 2004.

Suzuki H, Hayakawa S, Wada S, Okazaki E, Yamazawa M. Effect of solar drying on vitamin

D3 and provitamin D3 contents in fish meat. Journal of Agricultural and Food Chemistry 1988;

36: 803-806.

Taber L, Chiu CH, Whelan J. Assessment of the arachidonic acid content in foods commonly

consumed in the American diet. Lipids 1998; 33(12): 1151-1157.

Takeuchi A, Okano T, Teraoka S, Murakami Y, Kobayashi T. High-Performance Liquid-

Chromatographic Determination of Vitamin-D in Foods, Feeds and Pharmaceuticals by

Successive Use of Reversed-Phase and Straight-Phase Columns. Journal of Nutritional Science

and Vitaminology 1984; 30(1): 11-25.

Terry PD, Rohan TE, Wolk L. Intakes of fish and marine fatty acids and the risks of cancers of

the breast and prostate and of other hormone-related cancers: a review of the epidemiologic

evidence. American Journal of Clinical Nutrition 2003; 77(3): 532-543.

http://database.food.sugiyama-u.ac.jp/index_asia.php

http://database.food.sugiyama-u.ac.jp/index_asia.php

References

209

Thompson M, Sylvia G, Morrissey MT. Seafood traceability in the United States: Current

trends, system design, and potential applications. Comprehensive Reviews in Food Science and

Food Safety 2005; 4(1): 1-7.

Tidwell JH, Allan GL. Fish as food: aquaculture's contribution - Ecological and economic

impacts and contributions of fish farming and capture fisheries. Embo Reports 2001; 2(11):

958-963.

Tokudome Y, Imaeda N, Ikeda M, Kitagawa I, Fujiwara N, Tokudome S. Foods contributing

to absolute intake and variance in intake of fat, fatty acids and cholesterol in middle-aged

Japanese. Journal of Epidemiology 1999; 9(2): 78-90.

Tokudome Y, Kuriki K, Imaeda N, Ikeda M, Nagaya T, Fujiwara N, Sato J, Goto C, Kikuchi S,

Maki S, Tokudome S. Seasonal variation in consumption and plasma concentrations of fatty

acids in Japanese female dietitians. European Journal of Epidemiology 2003; 18(10): 945-953.

Tooze JA, Midthune D, Dodd KW, Freedman LS, Krebs-Smith SM, Subar AF, Guenther PM,

Carroll RJ, Kipnis V. A New Statistical Method for Estimating the Usual Intake of

Episodically Consumed Foods with Application to Their Distribution. Journal of the American

Dietetic Association 2006; 106(10): 1575-1587.

Törnkvist A, Aune M, Bjerselius R, Larsson L, Ankarberg E, Darnerud PO, Andersson A,

Bergkvist P, Andersson O, Glynn AW. The importance of specific and clear instructions for

sample preparation when analysing organochlorine compounds in fish. Organohalogen

Compounds 2005; 67: 1450-1453.

Tran NL, Barraj L, Smith K, Javier A, Burke TA. Combining Food Frequency and Survey

Data to Quantify Long-Term Dietary Exposure: A Methyl Mercury Case Study. Risk Analysis

2004; 24(1): 19-30.

Trondsen T, Scholderer J, Lund E, Eggen AE. Perceived barriers to consumption of fish

among Norwegian women. Appetite 2003; 41(3): 301-314.

References

210

Upritchard JE, Zeelenberg MJ, Huizinga H, Verschuren PM, Trautwein EA. Modern fat

technology: what is the potential for heart health ? Proceedings of the Nutrition Society 2005;

64(3): 379-386.

US Department of Agriculture and Agricultural Research Service. USDA National Nutrient

Database for Standard Reference. Release 16-1. Available at

http://www.nal.usda.gov/fnic/foodcomp/search/. 2005.

US Department of health and human services and US Environmental Protection Agency.

Mercury levels in commercial fish and shellfish. Available at

http://vm.cfsan.fda.gov/~frf/sea-mehg.html. 2004.

Van den Berg M, Birnbaum L, Bosveld ATC, Brunstrom B, Cook P, Feeley M, Giesy JP,

Hanberg A, Hasegawa R, Kennedy SW, Kubiak T, Larsen JC, van Leeuwen FXR, Liem AKD,

Nolt C, Peterson RE, Poellinger L, Safe S, Schrenk D, Tillitt D, Tysklind M, Younes M, Waern

F, Zacharewski T. Toxic equivalency factors (TEFs) for PCBs, PCDDs, PCDFs for humans

and wildlife. Environmental Health Perspectives 1998; 106(12): 775-792.

Van den Berg M, Peterson RE, Schrenk D. Human risk assessment and TEFs. Food Additives

and Contaminants 2000; 17(4): 347-358.

Van den Berg M, Birnbaum LS, Denison M, De Vito M, Farland W, Feeley M, Fiedler H,

Hakansson H, Hanberg A, Haws L, Rose M, Safe S, Schrenk D, Tohyama C, Tritscher A,

Tuomisto J, Tysklind M, Walker N, Peterson RE. The 2005 World Health Organization

Reevaluation of Human and Mammalian Toxic Equivalency Factors for Dioxins and Dioxin-

Like Compounds. Toxicological Sciences 2006; 93(2): 223-241.

Verbeke W, Sioen I, Pieniak Z, Van Camp J, De Henauw S. Consumer perception versus

scientific evidence about health benefits and safety risks from fish consumption. Public

Health Nutrition 2005; 8(4): 422-429.

Verbeke W, Vackier I. Individual determinants of fish consumption: application of the theory

of planned behaviour. Appetite 2005; 44(1): 67-82.

http://www.nal.usda.gov/fnic/foodcomp/search/

http://vm.cfsan.fda.gov/~frf/sea-mehg.html

References

211

Verbeke W, Vanhonacker F, Sioen I, Van Camp J, De Henauw S. Perceived importance of

sustainability and ethics related to fish: a consumer behaviour perspective. Ambio 2007a; In

Press.

Verbeke W, Sioen I, Brunsø K, De Henauw S, Van Camp J. Consumer perception versus

scientific evidence of farmed and wild fish: exploratory insights from Belgium. Aquaculture

International 2007b; 15(2): 121-136.

Visioli F, Rise P, Plasmati E, Pazzucconi F, Sirtori CR, Galli C. Very low intakes of N-3 fatty

acids incorporated into bovine milk reduce plasma triacylglycerol and increase HDL-

cholesterol concentrations in healthy subjects. Pharmacological Research 2000; 41(5): 571-576.

Voelker T, Kinney AT. Variations in the biosynthesis of seed-storage lipids. Annual Review of

Plant Physiology and Plant Molecular Biology 2001; 52: 335-361.

Vose D. Quantitative risk analysis. A guide to Monte Carlo simulation modeling. England:

John Wiley & sons, 1996.

Vrijens B, De Henauw S, Dewettinck K, Talloen W, Goeyens L, De Backer G, Willems JL.

Probabilistic intake assessment and body burden estimation of dioxin-like substances in

background conditions and during a short food contamination episode. Food Additives and

Contaminants 2002; 19(7): 687-700.

Warenwetgeving. Reglementering van samenstelling en benaming van voedingsmiddelen,

aanvulling nr. 38 (Regulation on the composition and appellation of food items, supplement

n° 38). Brugge: Die Keure, 1998.

Welch AA, Lund E, Amiano P, Dorronsoro M, Brustad M, Kumle M, Rodriguez M, Lasheras

C, Janzon L, Jansson J, Luben R, Spencer EA, Overvad K, Tjonneland A, Clavel-Chapelon F,

Linseisen J, Klipstein-Grobusch K, Benetou V, Zavitsanos X, Tumino R, Galasso R, Bueno-

De-Mesquita HB, Ocke MC, Charrondiere UR, Slimani N. Variability of fish consumption

within the 10 European countries participating in the European Investigation into Cancer

and Nutrition (EPIC) study. Public Health Nutrition 2002; 5(6B): 1273-1285.

References

212

Wenning RJ. Uncertainties and data needs in risk assessment of three commercial

polybrominated diphenyl ethers: probabilistic exposure analysis and comparison with

European Commission results. Chemosphere 2002; 46(5): 779-796.

Willett WC. Nutritional Epidemiology. Second ed. New York: Oxford University Press, 1998.

Williams CM, Burdge G. Long-chain n-3 PUFA: plant v. marine sources. Proceedings of the

Nutrition Society 2006; 65(1): 42-50.

Wolpowitz D, Gilchrest BA. The vitamin D questions: How much do you need and how

should you get it? Journal of the American Academy of Dermatology 2006; 54(2): 301-317.

World Health Organization. Guidelines for Predicting Dietary Intake of Pesticide Residues.

GEMS/Food in collaboration with the codex committee on pesticide residues. Document

WHO/FSF/FOS/97.7. Geneva: WHO, 1997.

World Health Organization. Food safety issues associated with products from aquaculture:

report of a joint FAO/NACA/WHO study group. Report number 883. Spain: WHO, 1999.

World Health Organization. Methodology for exposure assessment of contaminants and

toxins in food. WHO/SDE/PHE/FOS/00.5. Geneva: WHO, 2000.

World Health Organization. Diet, nutrition and the prevention of cardiovascular diseases.

WHO Technical report series, 916. Geneva: WHO, 2003.

World Health Organization. Safety evaluation of certain food additives and contaminants /

prepared by the fifty-seventh meeting of the Joint FAO/WHO Expert Committee on Food

Additives. WHO food additives series 48. Available at

http://www.who.int/ipcs/publications/jecfa/monographs/en/index.html. Geneva:

WHO, 2007.

Yaktine AL, Harrison GG, Lawrence RS. Reducing exposure to dioxins and related

compounds through foods in the next generation. Nutrition Reviews 2006; 64(9): 403-409.

http://www.who.int/ipcs/publications/jecfa/monographs/en/index.html

215

Summary

Pioneer research in the nineteen sixties and seventies indicated that the consumption of fish

was associated with a reduced risk for coronary heart disease in Greenland Eskimos. This

Eskimo population experienced a low mortality from coronary heart disease despite a diet

rich in fat and cholesterol. It was soon suggested that this could be related to the high

content of omega-3 fatty acids, typically present in marine foods. These Eskimo studies have

triggered a much broader and intensified research on the importance of omega-3 fatty acids

and seafood in the human diet.

This PhD-study is embedded in that research area and examines in the first place the intake

of omega-3 and other fatty acids by the Flemish and Belgian population. In a next step, the

question is raised whether seafood is a safe dietary source of these fatty acids and whether

the consumption conform with physiological needs induces any toxicological concern. The

latter is of importance since the favourable health perception is troubled by information

regarding the potential adverse health impact of chemical contaminants in marine foods,

occuring naturally or resulting from man-made processes. These conflicting facts form a

potential base for an important public health conflict between dietary recommendations on

the one hand and toxicological safety assurance on the other hand.

A first part of the PhD-thesis presents the results of an intake assessment of individual poly-

unsaturated fatty acids (PUFAs) in Flemish pre-school children, adolescents, and young

women and an intake assessment of vitamin D in Flemish adolescents. The following PUFAs

were considered: linoleic acid (LA), α-linolenic acid (LNA), arachidonic acid (AA),

eicosapentaenoic acid (EPA), docosapentaenoic acid (DPA), and docosahexaenoic acid

(DHA). First, the results make clear that a lower intake of LA and a higher intake of LNA

are necessary to decrease the LA/LNA ratio, since a lower LA/LNA ratio can help in the

prevention of some chronic diseases. This can be reached by a higher consumption of LNA

rich foods e.g. by replacement of omega-6 (n-6) rich oils by omega-3 (n-3) rich oils such as

linseed and rapeseed in food formulations. Second, dietary shifts are necessary to bridge the

gap between the intakes and the recommendations of long chain (LC) n-6 and n-3 PUFAs.

Regular replacement of meat products rich in saturated fatty acids (SFAs) by poultry meat is

a possible solution to increase the AA intake, which was evaluated to be very low for

216

Flemish pre-school children. Moreover, this study suggests that seafood and particularly

fatty fish consumption should be stimulated in all subgroups of the population since it is a

rich source of LC n-3 PUFAs of which the current intakes are evaluated to be far below the

recommendations. Furthermore, increased seafood consumption can lead to higher vitamin

D intake and can replace SFA-rich food items.

A next part of the thesis focuses on the nutritional-toxicological conflict related to seafood

consumption. This conflict arises from the fact that increased consumption of seafood will at

the same time increase the intake of environmental contaminants like methyl mercury

(MeHg), PCBs, and dioxin-like (dl) compounds (dl PCBs and PCDD/Fs). Therefore, a

methodology was developed, including the elaboration of databases and the development of

a software module in order to perform a probabilistic assessment of the simultaneous intake

of multiple compounds (LC n-3 PUFAs, vitamin D, iodine, (Me)Hg, PCBs, and dioxin-like

substances) via seafood consumption. Subsequently, the assessed intakes were evaluated to

determine: (1) if nutrient intakes reached the recommendations and (2) if contaminant

intakes did not exceed a level of toxicological concern.

Two different situations were studied:

1. Starting from the current seafood consumption data of two different subgroups of the Belgian

population (adolescents and adults).

Based on these consumption data, the simulation results predicted that the studied

populations did not reach a sufficiently high intake for the three nutrients under

consideration (LC n-3 PUFAs, vitamin D, iodine) taking into account only seafood

consumption. Regarding the contaminants, MeHg contamination of seafood on the Belgian

market did not seem to be an issue of toxicological concern. In contrast, the tolerable weekly

intake for dioxin-like compounds was exceeded in the case of high seafood consumers on the

basis of their seafood consumption only.

2. It was investigated whether the Belgian recommendation for LC n-3 PUFAs (i.e. 0.3 % of total

energy intake) could be reached through seafood consumption, without exceeding tolerable weekly

intakes of MeHg and dioxin-like compounds. Also the contribution of LC n-3 enriched margarines

was assessed.

The results indicated that the Belgian recommendation for EPA plus DHA can be reached by

consuming fatty fish twice a week, or by varying between lean and fatty fish minimally

217

three times a week. It was found that MeHg intake was not an issue of toxicological concern

for the Belgian population at this level of seafood consumption. The intake of dioxin-like

compounds from seafood only approximated the tolerable weekly intake when consuming

fatty fish three times a week or more, being a potential toxicological risk since also other food

items contribute to the daily intake of dioxin-like compounds.

Use of LC n-3 enriched margarine can further help to increase the LC n-3 PUFA intake, on

average by 159 mg/day. To conclude, combination of regular seafood consumption (twice a

week), with important contribution of fatty fish species, in combination with regular

consumption of LC n-3 enriched margarine can be advised to maximize LC n-3 intake.

There is, however, an important limitation related to this study, namely that contaminant

intake from other food sources was not taken into account. For MeHg this is not a problem

since exposure to MeHg occurs almost exclusively through consumption of seafood. In

contrast, consumption of other food items, mainly from animal origin, contributes to the

intake of dioxin-like compounds. This is an important fact that has to be taken into account

in further research.

219

Samenvatting

Baanbrekend onderzoek in de jaren zestig en zeventig van de vorige eeuw bracht aan het

licht dat de consumptie van vis het risico op hart- en vaatziekten deed dalen bij Eskimo’s in

Groenland. Deze Eskimobevolking vertoonde namelijk een lage mortaliteit ten gevolge van

hart- en vaatziekten ondanks een voedingspatroon rijk aan vet en cholesterol. Er werd

vermoed dat het hoge gehalte aan omega-3 vetzuren in mariene levensmiddelen hier een rol

in speelde. De studies bij deze Eskimobevolking gaven aanleiding tot verder intensief

onderzoek naar het belang van omega-3 vetzuren en vis en zeevruchten in het humane

voedingspatroon.

Ook dit doctoraatsonderzoek kadert binnen dat onderzoeksdomein en behandelt de inname

van omega-3 en andere vetzuren bij de Vlaamse en Belgische bevolking. Daarnaast werd ook

onderzocht of vis en zeevruchten een veilige voedingsbron zijn van deze omega-3 vetzuren

en welke toxicologische risico’s verbonden zijn aan visconsumptie die beantwoordt aan de

fysiologische noden voor deze vetzuren. Dit laatste is van belang omdat het gezonde imago

van vis en zeevruchten de laatste jaren verstoord wordt door negatieve boodschappen over

de aanwezigheid van chemische contaminanten in vis en zeevruchten. Deze tegenstrijdige

feiten vormen de basis voor een belangrijk conflict tussen voedingsaanbevelingen enerzijds

en toxicologische voorzorgsmaatregelen anderzijds.

Het eerste deel van deze thesis beschrijft de inname van de verschillende onverzadigde

vetzuren via het totale voedingspatroon voor drie Vlaamse bevolkingsgroepen: kleuters,

adolescenten en jonge vrouwen. De volgende vetzuren werden in rekening genomen:

linolzuur (LA), α-linoleenzuur (LNA), arachidonzuur (AA), eicosapentaeenzuur (EPA),

docosapentaeenzuur (DPA) en docosahexaeenzuur (DHA). Bovendien wordt de vitamine D

inname van adolescenten besproken. De resultaten maken duidelijk dat een lagere LA en

een hogere LNA inname nodig is om de verhouding van LA over LNA te doen dalen. Een

voedingspatroon met een lagere LA/LNA verhouding kan immers helpen bij de preventie

van sommige chronische ziekten. Praktisch kan dit worden bereikt door bijvoorbeeld omega-

6 rijke olieën te vervangen door omega-3 rijke olieën zoals lijnzaad- en koolzaadolie in

bepaalde samengestelde levensmiddelen zoals koekjes. Daarnaast zijn nog andere

verschuivingen in het voedingspatroon noodzakelijk om het verschil te herstellen tussen de

220

inname en de aanbevelingen van langketen omega-6 en omega-3 vetzuren. Om de AA-

inname die erg laag was bij de kleuters te verhogen, zou het goed zijn regelmatig

vleesproducten rijk aan verzadigde vetzuren te vervangen door gevogelte. Bovendien zou

voor alle bevolkingsgroepen de consumptie van vis en zeevruchten, en in het bijzonder

van vette vis, gestimuleerd moeten worden omdat deze voedingsmiddelengroep een

uitzonderlijk rijke bron is van langketen omega-3 vetzuren (EPA, DPA en DHA). De inname

van deze vetzuren ligt momenteel ver beneden de aanbevelingen. Daarenboven zal een

regelmatigere consumptie van vis en zeevruchten de inname van vitamine D verhogen en

zijn deze voedingsmiddelen een gezond alternatief voor voedingsmiddelen rijk aan

verzadigde vetzuren.

Het tweede deel van de thesis gaat over het nutritioneel-toxicologisch conflict verbonden aan

de consumptie van vis en zeevruchten. Dit conflict ontstaat door het feit dat een

regelmatigere visconsumptie tegelijkertijd de inname van chemische contaminanten zoals

methylkwik, PCBs en dioxines zal verhogen. Om dit van naderbij te kunnen bestuderen

werd een methodologie ontwikkeld waarbij databanken werden opgebouwd en waarbij een

software module werd ontwikkeld. Deze instrumenten lieten toe op probabilistische wijze de

gelijktijdige inname te berekenen van verschillende nutriënten en contaminanten (langketen

omega-3 vetzuren, vitamine D, jodium, methylkwik, PCBs en dioxineachtige componenten)

bij de consumptie van vis. Vervolgens werden de resultaten van deze innameschatting

gebruikt om te bepalen: (1) of de nutriëntinnames voldeden aan de aanbevelingen en (2) of

de contaminantinnames de toxicologische grenswaarden niet overschreden.

Binnen het doctoraatsonderzoek werden twee verschillende situaties bestudeerd:

1. Startend van het huidige patroon van vis- en zeevruchtenconsumptie van twee verschillende

Belgische bevolkingsgroepen (adolescenten en volwassenen).

Op basis van deze consumptiedata en de ontwikkelde methodologie werd voorspeld dat

beide populaties de aanbevelingen voor de verschillende nutriënten niet halen wanneer

enkel vis en zeevruchten als bron in rekening worden gebracht. Uit de evaluatie van de

inname van contaminanten kan worden besloten dat kwikcontaminatie van vis en

zeevruchten niet tot toxicologische risico’s leidt voor de Belgische bevolking. Daarentegen

werd de grenswaarde voor dioxineachtige componenten wel overschreden door mensen die

heel vaak vis consumeren (bv. drie tot vier porties vette vis per week). Bovendien werd deze

221

overschrijding reeds vastgesteld zonder het in rekening brengen van andere

voedingsmiddelen die ook dioxineachtige contaminanten bevatten.

2. Er werd onderzocht of het mogelijk is de Belgische aanbeveling voor langketen omega-3 vetzuren

(EPA en DHA) te halen louter op basis van vis- en zeevruchtenconsumptie, zonder dat de

grenswaarden voor methylkwik en dioxineachtige componenten worden overschreden. Ook werd

het mogelijke belang van omega-3 verrijkte margarines onderzocht.

De resultaten van dit onderzoeksluik tonen aan dat de Belgische aanbeveling voor EPA en

DHA gehaald kan worden wanneer twee keer per week vette vis geconsumeerd wordt. De

aanbeveling kan ook worden gehaald door te variëren tussen vette en magere vis, maar dan

moet vis drie keer per week op het menu staan. Hierbij is het wel belangrijk te vermelden

dat de inname van dioxineachtige componenten de grenswaarde bereikt wanneer vette vis

drie keer of meer per week op het menu staat. Wetende dat er in ons dagelijks

voedingspatroon nog andere bronnen zijn van deze dioxineachtige contaminanten, is dit dus

een belangrijk aandachtspunt. Het gebruik van margarine verrijkt met langketen omega-3

vetzuren kan verder helpen de inname van omega-3 vetzuren te verhogen.

Tot slot is het belangrijk nog eens te herhalen dat in deze doctoraatsthesis de

contaminantinname enkel werd beschouwd via de consumptie van vis en zeevruchten.

Andere mogelijke bronnen werden dus buiten beschouwing gelaten. Deze beperking heeft

geen grote invloed op de innameschatting van methylkwik, aangezien er weinig andere

bronnen van methylkwik in ons voedingspatroon aanwezig zijn. Het humane

voedingspatroon bevat echter wel nog andere bronnen van dioxineachtige contaminanten,

voornamelijk andere levensmiddelen van dierlijke oorsprong. Bij verder onderzoek is het

dan ook erg belangrijk deze mee in rekening te brengen.

223

Dankwoord / Acknowledgments

Aan het einde van dit werk durf ik zeggen trots te zijn op wat ik heb bereikt. Ik heb het werk

heel graag gedaan en ik heb genoten van de afgelegde weg. Uiteraard wil ik heel wat

mensen danken die me op allerlei manieren hebben begeleid, gestimuleerd,

geënthousiasmeerd en gemotiveerd.

In de eerste plaats wil ik mijn promotoren Prof. Dr. Stefaan De Henauw en Prof. Dr. ir. John

Van Camp bedanken voor hun wetenschappelijke ondersteuning en hun geloof in mij en

mijn werk. Als kritische geest stond ik wel vaak klaar met een tegen-woord, maar jullie

namen altijd de tijd om naar me te luisteren en te overleggen. Ook wil ik jullie bedanken

voor de vele kansen die ik kreeg om zowel in binnen- als buitenland verschillende

workshops, colloquia, congressen en cursussen te volgen. Stefaan, bedankt voor alles! Ik durf

zeggen dat we op de werkvloer samen gegroeid zijn. Het spreekwoord ‘zonder wrijving

geen warmte’ is vermoedelijk wel van toepassing. Heel tevreden terugkijkend wil ik je

bedanken voor de warme werkomgeving en de goede band die we hebben opgebouwd.

John, voor jou een woord van dank voor al je input, je nauwgezette opvolging van mijn werk

en de goede samenwerking.

De twee andere leden van mijn begeleidingscommissie, Prof. Dr. Jan Willems en Prof. Dr.

Guy De Backer, die beiden mee aan de bakermat stonden van het wetenschappelijke project

omtrent visconsumptie, geef ik graag een speciaal woord van dank. Professor Willems, dank

om zelfs na je emeritaat ons nog wekelijks te komen vervoegen en de tijd uit te trekken om

mijn werk kritisch na te lezen en te voorzien van de nodige opbouwende commentaar.

Daarnaast wil ik nog een aantal mensen bedanken met wie ik heel graag samen werkte

binnen het project omtrent de risico’s en baten van visconsumptie: Dr. ir. Frederik Verdonck,

Nicky Van Thuyne, Prof. Dr. ir. Peter Vanrolleghem, Prof. Dr. ir. Wim Verbeke, Dr. ir. Marc

Raemaekers en ir. Christine Hoefkens. Frederik, bedankt voor de vele statistische

ondersteuning, het geduldig uitleggen en de vele keren dat je me een hart onder de riem

stak. Nicky, Frederik en Peter, zonder jullie hulp bij het ontwikkelen van de nodige software,

had ik nooit zo snel deze doctoraatsweg kunnen bewandelen. Wim, bedankt voor de heel

224

warme en vlotte samenwerking. Ik vond het steeds erg aangenaam en verrijkend om samen

te discussiëren over het projectwerk en de verschillende publicaties die daaruit zijn

voortgevloeid. Marc, de tijd die we samen doorbrachten op het Departement Zeevisserij in

Oostende aan het begin van mijn doctoraatswerk ben ik nog niet vergeten. Bedankt daarvoor

en vooral ook voor de vele ‘trucjes’ die je me leerde om sneller en efficiënter in Excel® te

kunnen werken. Ik maak er nog dagelijks gebruikt van. Christine, als thesisstudente heb je

een mooie bijdrage geleverd aan mijn doctoraatsonderzoek.

I want to thank Prof. Dr. J.M. Kaufman, Prof. Dr. J. Van de Voorde, Prof. Dr. C. Van

Peteghem, Prof. Dr. ir. G. Vansant and Dr. J.L. Volatier for being part of the examination

committee. Aussi un grand merci pour Dr. Volatier, Dr. Leblanc et tous les collègues de

l’AFSSA pour m’avoir accueilli à Paris.

Uiteraard wil ik ook al mijn collega’s bedanken, zowel op het UZ als op het boerenkot. Een

aantal collega’s verdienen een ereplaatsje. Maaike, mijn bureaugenoot, mijn luisterend oor,

mijn redder-in-computer-nood, mijn congres-partner, mijn medevismadam, … Ik heb

ontzettend veel aan je gehad op allerlei vlakken en wil je daarvoor heel erg danken. Alle

huidige en vorige VVV-collega’s wil ik bedanken voor de erg warme en open samenwerking.

Anne Opsomer, jou wil ik bedanken voor de vele administratieve ondersteuning en je

enthousiaste interesse. Els en Sofie, bedankt voor de vele babbels tussendoor als ik even

nood had aan ontspanning of om mijn hart te luchten. En een heel warm woord van dank

aan alle collega’s die ik hier niet met naam noem maar wel in gedachten heb.

Voor ik mijn werkomgeving verlaat wil ik uiteraard ook het IWT-Vlaanderen en het Federaal

Wetenschapsbeleid danken voor de financiële ondersteuning.

Wanneer ik het over de persoonlijke boeg gooi, komt Hans uiteraard op de eerste plaats.

Mijn man, mijn grote liefde, mijn ‘reason to live’. Ik wil je danken voor het samen onderweg

zijn, het samen geloven in wat voor en achter ons ligt en ook voor alles wat ik van je leerde.

Je leerde me relativeren, genieten en tevreden zijn. Zaken die ik ook op het werk goed kon

gebruiken.

225

Daarnaast wil ik heel graag mijn mama bedanken voor de vele grote en kleine momenten

samen. Ook mijn papa, Frederik, Magalie, Benoit, en alle andere dichte en verre familie- en

schoonfamilieleden, bedankt voor het samen onder weg zijn!

En dan die ontzettend fijne, grote en kleine vriendschappen om me heen. Hier laat ik het na

om namen te noemen, want ik wil niemand vergeten en ook geen volgorde geven. Maar zij

die zich aangesproken voelen, voelen het goed. Beste lieve vrienden, bedankt omdat jullie er

zijn, in mij geloven, we samen kunnen lachen, genieten en plezier maken en we naar elkaar

kunnen luisteren. Eén vriend verdient echter speciale aandacht in dit dankwoord. Gert,

bedankt voor je onuitputtelijk geloof en enthousiasme. Steeds opnieuw bevestigde je mij in

mijn werk, hier en daar hielp je me bij een presentatie en telkens weer was je geïnteresseerd

om te luisteren naar wat ik nu weer had bekokstoofd. Bedankt!

Isabelle Sioen

Oktober 2007

227

About the author

Isabelle Sioen was born on July 15, 1980 in Ghent (Belgium). She finished secondary school in

Latin and mathematics at the ‘Sint-Bavohumaniora’ in Ghent in 1998. The same year, she

started the study of bioscience engineering at the Ghent University. In 2001, she studied

during one semester at the Department of Science and Technologies of the Food Industry at

the ISIM (Institut des Sciences de l’Ingénieur de Montpellier II) in France with an Erasmus

program. During the last year of her master studies, she conducted three months of

fieldwork in the Houndé District in Burkina Faso for her dissertation entitled ‘Evaluation of

the nutritional status and dietary habits of pregnant women in rural Burkina Faso’. She

obtained her MSc degree in Bioscience Engineering, option chemistry, in 2003.

From November 2003 until February 2004, she worked as a scientific staff member at the

Research and Information Centre for Consumer Organizations in Brussels.

In March 2004, she started working at the Ghent University as a PhD student on a project

financed by the Belgian federal government, half of the time at the Department of Public

Health of the Faculty of Medicine and Health Sciences and half of the time at the Department

of Food Safety and Food Quality at the Faculty of Bioscience Engineering. In January 2005,

she obtained an IWT-scholarship and she continued her PhD-work. During this period, she

followed different international courses, for instance the ‘EuroFIR Course on Production of

Food Composition Data in Nutrition’ in Bratislava (Slovak Republic) and the ‘Summer

School EU Basics in Public Health Nutrition’ in Giessen (Germany). From February 10, 2007

to May 10, 2007 she was hosted at the French Food Safety Agency (AFSSA, Agence Française

de Sécurité Sanitaire des Aliments) in Paris (France) for a scientific mission in relation to her

PhD-work.

She participated in many international congresses and workshops where she presented her

work by means of posters or oral communications. At the 1st International Congress on Food

Safety with the topic ‘Nutrition and Food Safety: Evaluation of Benefits and Risks’ held in

June 2006, in Budapest (Hungary), she won the Award for Best Student Poster with her

poster ‘Probabilistic risk-benefit analysis regarding seafood consumption’. At the 10th

European Nutrition Conference held in July 2007 in Paris (France), she won the Young

228

Scientist FENS Award for Oral Communication, with her presentation ‘Dietary intake and

food source of omega-3 and omega-6 poly-unsaturated fatty acids in the Belgian population’.

In the summer of 2007, she won the Prof. Dr. G. Verdonk Award for Dietetics presented by

the Belgian Royal Academy for Medicine (period 2003-2006) with her work entitled

‘Evaluation of benefits and risks related to seafood consumption’.

229

Publications of the author

Articles in international peer-reviewed journals included in Science Citation Index (A1) Verbeke W, Sioen I, Pieniak Z, Van Camp J, De Henauw S. Consumer perception versus scientific evidence about health benefits and safety risks from fish consumption. Public Health Nutrition 2005; 8(4):422-429. Sioen I, Haak L, Raes K, Hermans C, De Henauw S, De Smet S, Van Camp J. Effects of pan-frying in margarine and olive oil on the fatty acid composition of cod and salmon. Food Chemistry 2006; 98(4):609-617. Sioen I, Pynaert I, Matthys C, De Backer G, Van Camp J, De Henauw S. Dietary intakes and food sources of fat fatty acids for Belgian women, focused on n-6 and n-3 polyunsaturated fatty acids. Lipids 2006; 41(5):415-422. Haak L, Sioen I, Raes K, Van Camp J, De Smet S. Effect of pan frying in different culinary fats on the fatty acid profile of pork. Food Chemistry 2007; 102(3):857-864. Sioen I, Van Camp J, Verdonck F, Van Thuyne N, Willems J, De Henauw S. How to use secondary data on seafood contamination for probabilistic exposure assessment purposes? Main problems and potential solutions. Human and Ecological Risk Assessment 2007; 13:632-657. Verbeke W, Sioen I, Brunsø K, De Henauw S, Van Camp J. Consumer perception versus scientific evidence of farmed and wild fish: exploratory insights from Belgium. Aquaculture International 2007; 15: 121-136. Sioen I, De Henauw S, Verdonck F, Van Thuyne N, Van Camp J. Development of a nutrient database and distributions for use in a probabilistic risk-benefit analysis of human fish consumption. Journal of Food Composition and Analysis 2007; 20(8): 662-670. Sioen I, Huybrechts I, Verbeke W, Van Camp J, De Henauw S. n-6 and n-3 PUFA intakes of pre-school children in Flanders, Belgium. British Journal of Nutrition 2007; 98(4): 819-825. Sioen I, Matthys C, De Backer G, Van Camp J, De Henauw S. Importance of seafood as nutrient source in the diet of Belgian adolescents. Journal of Human Nutrition and Dietetics 2007; In Press. Bilau M, Sioen I, Matthys C, Goemans G, Belpaire C, De Vocht A, De Henauw S, Willems JL Probabilistic approach to polychlorinated biphenyl (PCB) exposure through eel consumption in recreational fishermen vs. the general population. Food Additives and Contaminants 2007; In Press. Available online doi: 10.1080/02652030701459848 Antonjevic B, Matthys C, Sioen I, Bilau M, Van Camp J, Willems J, De Henauw S. Simulated impact of a fish based shift in the population n-3 fatty acid intake on exposure to dioxins and dioxin-like compounds. Food and Chemical Toxicology 2007; In Press. Available online doi:10.1016/j.fct.2007.06.003.

230

Verbeke W, Vanhonacker F, Sioen I, Van Camp J, De Henauw S. Perceived importance of sustainability and ethics related to fish: a consumer behavior perspective. Ambio 2007; In Press. Sioen I, Bilau M, Verdonck F, Verbeke W, Willems J, De Henauw S, Van Camp J. Probabilistic intake assessment of polybrominated diphenyl ethers and omega-3 fatty acids through fish consumption. Molecular Nutrition and Food Research 2008; In Press. Sioen I, Verbeke W, De Henauw S, Parmentier K, Raemaekers M, Willems J, Van Camp J. Traceability of seafood on the Belgian market. Fisheries Research. Submitted. Sioen I, Van Camp J, Verdonck F, Verbeke W, Vanhonacker F, Willems J, De Henauw S. Probabilistic intake assessment of multiple compounds as a tool to quantify the nutritional-toxicological conflict related to seafood consumption. Chemosphere. Submitted. Sioen I, De Henauw S, Verbeke W, Verdonck F, Willems J, Van Camp J. Fish consumption is a safe solution to increase the intake of long chain omega-3 fatty acids. Public Health Nutrition. Submitted. Articles in international peer-reviewed journals not included in Science Citation Index (A2) Sioen I, Bilau M, Van Camp J, De Henauw S. Encountered problems when using published fish contamination data in a probabilistic intake assessment. Organohalogen Compounds 2005; 67:2473-2475. Bilau M, Sioen I, Matthys C, Goemans G, Belpaire C, De Vocht A, De Henauw S. PCB exposure through eel consumption in sport fishers as compared to the general population: a probabilistic approach. Organohalogen Compounds 2005; 67:1737-1740. Sioen I, Van Camp J, Verdonck F, Van Thuyne N, Vanrolleghem PA, Vanhonacker F, Verbeke W, De Henauw S. Risk-benefit analysis regarding seafood consumption: a tool for combined intake assessment. Organohalogen Compounds 2006; 68:379-382. Sioen I, Bilau M, De Knuydt M, Van Camp J, De Henauw S. Intake assessment of polybrominated flame retardants by seafood consumption. Organohalogen Compounds 2006; 68:2177-2180. Verdonck FAM, Sioen I, Baert K, Van Thuyne N, Bilau M, Matthys C, De Henauw S, De Meulenaer B, Devlieghere F, Van Camp J, Vanrolleghem PA, Van Sprang P, Verbeke W, Willems J. Uncertainty and Variability Modelling of Chemical Exposure through Food. Archives of Public Health 2006; 64:159-173. Sioen I, Van Camp J, Verdonck F, Vanhonacker F, Verbeke W, De Henauw S. Nutritional-toxicological conflict of fish consumption: a tool for combined intake assessment. Communication in Agricultural and Applied Biological Sciences 2006; 71(1):263-266.

231

Articles in national journals, not under A1 and A2 Sioen I. Omega-3 vetzuren: mag het iets meer zijn? De Eetbrief. UGent Nieuwsbrief over Gezond en Lekker Eten 2006; 145: 3-4. Sioen I, Van Camp J, De Henauw S. De inname van omega-6 en omega-3 vetzuren bij de Belgische bevolking. Tijdschrift voor Voeding en Diëtetiek 2007; 1:2-6. Sioen I, De Henauw S, Van Camp J. Evaluation of benefits and risks related to seafood consumption. Verhandelingen van de Koninklijke Academie voor Geneeskunde van België 2007. In Press. Book De Henauw S, Van Camp J, De Meulenaere B, Andjelkovic M, Bilau M, Huybrechts I, Huybregts L, Matthys C, Mestdagh F, Opsomer A, Pynaert I, Sioen I, Vercruysse L. Een overzicht van de relatie voeding en gezondheid, en van functionele componenten in levensmiddelen. Flanders Food, Fevia (ed), Leuven (Belgium): LannooCampus, 2006. Chapter in book Vanhonacker F, Verbeke W, Sioen I. Consumer perception about ethical and sustainability issues of fish. In: Ethics and the politics of food. Wageningen, The Netherlands: Wageningen Academic Publishers, 2006: 464-469 Report De Henauw S, Willems J, Sioen I, Van Camp J, Verbeke W, De Lange I, Vanhonacker F, Cooreman K, Raemaekers M, Parmentier K, Vanrolleghem P, Verdonck F, Vanthuyne N. Integrated evaluation of marine food items: nutritional value, safety and consumer perception. Final Report CP/56 published by the Belgian Science Policy, Brussels, July 2006. Abstracts Lanou H, Roberfroid D, Ki J-P, Koudogbo V, Sia D, Henry M-C, Meda N, Korgo P, Sioen I, Termote C, Kolsteren P. Le project MISAME. Prévention du retard de croissance intra-untérine dans le district de Houndé, Burkina Faso. International Symposium, Research in Applied Nutrition in Developing Countries: Challenges and Expectations, Brussels, Belgium, December 2004. Sioen I, Hermans C, De Henauw S, Van Camp J. Effects of pan frying in margarine and olive oil on the fatty acid content and composition of salmon and cod. 7e Voedings- en Gezondheidscongres, Brussels, Belgium, November 2004. Pieniak Z, Verbeke W, De Lange I, Sioen I, Van Camp J, De Henauw S. Determinants of fish consumption: Application of the Theory of Planned Behaviour. 7e Voedings- en Gezondheidscongres, Brussels, Belgium, November 2004.

232

De Lange I, Sioen I, Pieniak Z, Verbeke W, Van Camp J, De Henauw S. Health benefits and safety risks from fish consumption: consumer perception versus scientific evidence. 7e Voedings- en Gezondheidscongres, Brussels, Belgium, November 2004. Sioen I, Haak L, Raes K, Hermans C, De Henauw S, De Smet S. Van Camp J. Effects of pan frying in margarine and olive oil on the fatty acid content and composition of salmon and cod. KVCV-symposium "Voedselchemie in Vlaanderen V: Trends in de Levensmiddelenanalyse", Ghent, Belgium, March 2005. Sioen I, Antonijevic B, Matthys C, Bilau M, De Henauw S, Willems J. Probabilistic intake assessment (PIA): exposure to contaminants via food. Workshop ‘Probabilistic risk assessment: current developments and applications for environmental assessment and management’. Society for Risk Analysis and Interstate Technology Regulatory Council Workshop. Michigan State University, East Lansing, Michigan, USA, March 2005. Sioen I, De Henauw S, Van Camp J. Building up a nutrient database for a probabilistic risk-benefit analysis regarding fish consumption. 6th International Food Data Conference, Pretoria, South Africa, September 2005. Sioen I, Haak L, Raes K, Hermans C, De Henauw S, De Smet S. Van Camp J. Effects of pan frying in margarine and olive oil on the fatty acid composition of cod and salmon. 35th WEFTA meeting, Antwerp, Belgium, September 2005. Sioen I, Raemaekers M, De Henauw S, Van Camp J, Willems J. Building up databases to use a probabilistic risk-benefit analysis regarding fish consumption. 35th WEFTA meeting, Antwerp, Belgium, September 2005. Raemaekers M, Sioen I, De Henauw S, Van Camp J, Willems J. The origin of seafood products on the Belgian market. 35th WEFTA meeting, Antwerp, Belgium, September 2005. De Lange I, Verbeke W, Sioen I, Van Camp J, De Henauw S, Brunsø K. Consumer perception versus scientific evidence of farmed and wild fish: exploratory insights in Belgium. 35th WEFTA meeting, Antwerp, Belgium, September 2005. Sioen I, Matthys C, Pynaert I, Van Camp J, De Henauw S. Dietary intake and food source of poly unsaturated fatty acids in Belgian women. 8e Voedings- en Gezondheidscongres, Brussels, Belgium, November 2005. Bilau M, Sioen I, Matthys C, De Vocht A, Willems JL, De Henauw S. iPCB exposure through eel consumption (sport fishers versus general population): a probabilistic approach. 8e Voedings- en Gezondheidscongres, Brussels, Belgium, November 2005. Sioen I, Verdonck F, Van Thuyne N, Vanrolleghem PA, De Henauw S, Van Camp J. Probabilistic risk-benefit analysis regarding seafood consumption. SETAC Europe 16th Annual Meeting, The Hague, The Netherlands, May 2006. Sioen I, Van Camp J, De Henauw S. Dietary intake and food sources of polyunsaturated fatty acids in Belgian women. Europrevent 2006, Athens, Greece, 11-13 May 2006. European Journal of Cardiovascular Prevention & Rehabilitation 2006; 13(suppl1):S30.

The Nutritional-Toxicological Conflict related to Seafood ...fatty acids. Next, an introduction to...

Documents

Transcript of The Nutritional-Toxicological Conflict related to Seafood ...fatty acids. Next, an introduction to...